Fed Is Ignoring Actual Inflation Data

Policy makers may be relying on the wrong model as they push for a December rate hike.

Former Fed Governor Daniel Tarullo isn't so sure.

Photographer: Ron Antonelli/Bloomberg

Federal Reserve policy makers tend to believe temporary shocks account for the persistent undershooting of the inflation target. But there's a more disturbing possibility: Central bankers might just be using a broken model of inflation. Given that risk, they should pay attention to actual inflation and lean toward passing on a December rate hike. Nonetheless, they are prepared to move forward. The dismissal of actual data given these very real concerns about the forecasting accuracy of the Fed’s models could place us in a more dangerous stage of the business cycle.

The Fed uses a monetary policy rule as a guide to setting short-term rates. Policy makers focus on Taylor-type rules, which make policy a function of the neutral interest rate (otherwise known as “r-star”), the central bank’s inflation target, deviations in actual inflation relative to the target, and deviations of unemployment from its natural rate.

Former Fed Governor Daniel Tarullo highlighted the challenges of this approach in a recent paper. He pointed out that many of the key variables in a Taylor rule are unobserved and need to be estimated, creating a great deal of policy uncertainty. An estimate of the natural rate of unemployment in excess of its true level, for example, might result in an excessively tight monetary policy.

Tarullo takes particular issue with the Fed’s inflation model. Policy makers employ an expectations-augmented Phillips curve model, in which inflation is a function of expectations and the tightness of the labor market. As is well known, this approach has fallen short in recent years. The Fed repeatedly forecast inflation would return to its 2 percent target, only to be disappointed that the level remains consistently below that level.

One explanation for the inflation shortfall is a weaker relationship between unemployment and inflation. Tarullo, however, zeros in on the problems surrounding the most important element of the Fed’s inflation model, inflation expectations. Unlike other unobserved variables, inflation expectations

... are not simply a factor or parameter that helps determine the path within which the economy is likely to head with a given monetary policy. Instead, they are believed to exert a direct effect upon the inflation rate in the economy and, most importantly for present purposes, to be significantly within the control of the central bank. 

According to central bankers, inflation expectations not only set actual inflation, but policy makers control those expectations. Thus “anchoring” expectations around the chosen target takes on a critical role for the Fed.

The emphasis on inflation expectations is misguided, Tarullo finds. The central bank lacks critical elements necessary to implement a model dependent on inflation expectations. Policy makers have no widely accepted metric for gauging inflation expectations, no clear understanding of the link between expectations and actual prices, and do not know how and why inflation expectations change.

Tarullo draws an unsettling conclusion:

After eight years at the Fed (actually, well before that) my conclusion was that there is no well-elaborated and empirically grounded theory that explains contemporary inflation dynamics in a way useful to real-time policy making.  

Tarullo sees three implications for policy. First, the Fed should pay more attention to observables than to unobservables. Second, given a lack of a useful inflation model, the bank should not tie policy, either implicitly or explicitly, to specific rules. This is obviously a retort to “audit the Fed”-type bills under consideration in Congress. And, finally, the Fed faces a period when the Phillips curve is not a workable policy guide.

If Tarullo is correct, the Fed is on shaky grounds with a December rate hike. Actual inflation is well below the year-ago forecast, which should induce the Fed to pass on December in favor of a faster pace of rate hikes next year, if necessary. But although policy makers increasingly question elements of their inflation model, most are not ready to abandon the Phillips curve framework entirely. Moreover, they hesitate to consider a faster pace of rate hikes, believing this to be a path to near-certain recession. Consequently, central bankers remain committed to policy gradualism.

Given Tarullo’s analysis, market participants should be aware of some fairly significant monetary policy dangers over the next year. One risk is that the Fed mistakenly adheres to a broken Phillips curve framework and continues to tighten policy despite low inflation. This risks excessive tightening, unnecessarily slowing activity and possibly triggering the recession gradualism is meant to avoid.

What concerns me most, however, is that for the Fed’s gradualism approach to work, policy makers need to resist stepping up the pace of rate increases should inflation accelerate. After all, they already tightened in advance of that quickening. I am worried that they will lose sight of the lags in monetary policy and, upon seeing inflation accelerate, conclude that their Phillips curve approach was correct all along and that they are falling behind the curve. This would almost certainly be a recipe for recession.

    To contact the author of this story:
    Tim Duy at

    To contact the editor responsible for this story:
    Max Berley at

    Before it's here, it's on the Bloomberg Terminal.