Broken Phillips Curve a Symptom of Lower U.S. Inflation Expectations

We believe the Phillips curve framework is doing a poor job at forecasting inflation.

Is the Phillips curve breaking down? Tight labor markets have historically tended to place upward pressure on wages and inflation. And the Phillips curve, a theory embodying the inverse relationship between inflation and unemployment, has been the cornerstone of modern monetary policy.

Yet U.S. CPI and wage inflation remain modest today, despite a nearly eight-year U.S. economic expansion and an unemployment rate that has already dropped below the Federal Reserve’s estimated “natural rate of unemployment” (the “equilibrium” long-run level of unemployment, often used interchangeably with the “non-accelerating inflation rate of unemployment,” or NAIRU). This casts doubt on the importance of labor market slack as a driver of inflation.

It’s important to note that future realized inflation depends on both historically realized inflation and inflation expectations. Phillips curve believers – unable to explain the low inflation rate in the current setting – might argue that this is because U.S. inflation expectations are very well anchored. Indeed, studies have documented that the weight of long-term expectations in the Phillips curve has risen steadily since the mid-1980s, while the slope of the Phillips curve has substantially declined, and the curve today could be flat.1

Make no mistake, a flat Phillips curve would be a welcome development for the Fed, if inflation were on target. The reason is fairly straightforward: If people expect long-term inflation to be stable, wages won’t react as much to near-term changes in labor market slack. In this scenario, if a shock were to push inflation up, then anchored expectations would push it back down without the need for monetary tightening (and vice versa).2  

But here’s the hitch: Even though U.S. inflation expectations are supposed to be very well anchored, no one can tell where.

An elusive anchor

So what are some of the possibilities?

Looking at core personal consumption expenditure (PCE) inflation (see Figure 1), the Fed’s preferred benchmark, one could speculate that the U.S. is well anchored at 1.5% – not the 2% Fed target – just as Japan and the eurozone appear solidly anchored below target at 0% and 1%, respectively. Here the Fed usually points to professional surveys of inflation expectations, which tend to show somewhat higher values. Yet professional forecasters’ longer-term expectations have exceeded realized inflation almost 70% of the time since the early 1990s (when records began).3

Figure 1 is a scatter plot of the unemployment rate versus core personal consumption expenditure (PCE) inflation over the period 2000 to 2017. The average of all of the plots forms a downward sloping line, showing that high unemployment rates tend to correspond with relatively low core PCE, and vice versa. The plots indicate that core PCE, shown on the X-axis, ranges between roughly 1% and 2.5%, while the unemployment rate ranges between roughly 4% and 10%.

Similarly, the Fed has been systematically discounting low market-based inflation expectations, attributing them to either reduced liquidity in the Treasury Inflation-Protected Securities (TIPS) market or to changes in the inflation risk premium. We disagree with the Fed here. As we’ve pointed out in the past, the distribution of potential inflation outcomes – and market perceptions of the likelihood of downside or upside misses – is at least as relevant as the mode, if not more. The latest research from the European Central Bank (ECB) suggests that a negative inflation risk premium is associated with strong deflation risk, rather than low inflation uncertainty.4 In essence, the market assigns a higher probability to the Fed’s missing to the downside.

To further complicate matters, prolonged deviations of inflation from the target may eventually de-anchor expectations, and it’s hard for anyone to know what margin monetary policy has before this happens.

Meanwhile, the unemployment rate fell below the Fed’s “natural rate” estimate in December 2016, which should (in theory) intensify upward pressure on prices (see Figure 2). Instead, remarkably weak CPI prints in March and April brought the Fed’s preferred measure of market-based long-term inflation expectations down to 1.9%, around half a percentage point below the Fed’s target in CPI terms (see Figure 3). The April CPI report reflected broad-based weakness across a range of core goods and services, prompting us to revise our 2017 year-end PCE forecast to 1.5%, versus the Federal Open Market Committee (FOMC) median projection of 1.9% and its 2% PCE target.

Figure 2 is a line graph showing the unemployment rate and the CBO NAIRU (non-accelerating inflation rate of unemployment) estimate, from September 2012 to 30 April 2017. For most of the period, the unemployment rate is above the CBO NAIRU estimate, starting at around 8%, and declining to about 4.5% in 2017. The CBO NAIRU estimate starts at just above 5% and is almost flat, declining slightly to just beneath 5% during the period. The unemployment rate falls to the level of the CBO NAIRU estimate by mid-2016, and falls below it in early 2017.

Figure 3 is a line graph showing the market-based inflation expectations from 1999 to May 2017. For most of the period, the Fed’s 5-year, 5-year forward breakeven inflation is above that of the CPI (consumer price index) target, which is shown by a solid horizontal line at around 2.4%. Inflation expectations, shown on the Y-axis, start at about 1.6% in 1999, then rise to a higher range, above the CPI target of 2.4%, roughly between 2.4% and 3.5%. Yet in 2014, the expectation line falls below the CPI target of 2.4%, and continues a downward trend, ending at around 1.8% in 2017. Inflation expectations show a bottom of about 1.4% in 2016.

How low will the NAIRU go?

The natural rate of unemployment is also something of a moving target. Only a few years ago, the NAIRU – which is itself a construct, and thus not directly observable – was considered to be about 6%. But with inflation nowhere to be seen, the Fed gradually lowered it all the way down to 4.7%. Since the NAIRU estimate is derived from a variant of the Phillips curve relationship, it seems logical to acknowledge the risk that it will continue to fall along with the unemployment rate.

This introduces another problem: While everyone accepts that the U.S. economy is closing the unemployment gap, no one really knows how close we are – or what the true natural rate of unemployment actually is.

Key takeaways

Softening inflation calls into question the Fed’s presumed plan to hike interest rates in June, notwithstanding that the unemployment rate has been falling faster than the Fed had originally anticipated. In this context, we think a hawkish bias appears inconsistent with a data-dependent approach.

The Fed has recently maintained that its 2% inflation target is symmetric, but hiking rates when inflation is below target (and slowing) could signal to investors that monetary policy will more likely than not miss the target on the downside. We think pausing here, waiting for inflation to reach 2% and eventually resuming tightening once it can better pinpoint the NAIRU could be a better approach. If inflation expectations continue to lie below the target, this would only undermine the Fed’s credibility in carrying out its price stability mandate, and the market is no longer giving the Fed the benefit of the doubt on this one.

In sum, we believe the Phillips curve framework is doing a poor job at forecasting inflation, even after tweaking the two main inputs: inflation expectations and (to a lesser extent) the NAIRU. Inflation expectations seem to be well anchored, but no one knows where. And the NAIRU could, at least theoretically, drop all the way down to 0% before inflation accelerates meaningfully.

In the meantime, the Fed faces a stark choice: continue to tighten policy preemptively, or listen to what the data and the markets have to say.

1 Blanchard, Olivier. Peterson Institute for International Economics Policy Brief, “The US Phillips Curve: Back to the 60s?” January 2016.
2 Blanchard, Olivier, Eugenio Cerutti and Lawrence Summers. IMF Working Paper, “Inflation and Activity: Two Explorations and their Monetary Policy Implications,” November 2015.
3 Federal Reserve Bank of Philadelphia’s “Survey of Professional Forecasters” on projections for the 10-year CPI inflation rate.
4 Camba-Mendez, Gonzalo and Thomas Werner. ECB Working Paper Series, “The inflation risk premium in the post-Lehman period,” March 2017.
The Author

Spyros Michas

Portfolio Manager, Real Return

View Profile

Latest Insights



This material contains the opinions of the manager and such opinions are subject to change without notice. This material has been distributed for informational purposes only and should not be considered as investment advice or a recommendation of any particular security, strategy or investment product. Information contained herein has been obtained from sources believed to be reliable, but not guaranteed. No part of this material may be reproduced in any form, or referred to in any other publication, without express written permission. PIMCO is a trademark of Allianz Asset Management of America L.P. in the United States and throughout the world. ©2017, PIMCO.