Data Dependence, Broadly Defined – Implications Of Last Week’s Fed Meeting by ColumbiaManagement

  • Last week’s FOMC meeting was the third largest dovish surprise in the QE era, only bested by the original QE1 announcement and the September 2013 “no taper” decision.
  • We continue to expect the FOMC to hike rates in September, and the pace of rate hikes thereafter should be faster than markets are currently pricing.
  • In this environment, the U.S. dollar can continue to outperform and U.S. rates underperform, even with a Fed moving a bit more gingerly towards liftoff.

At last week’s meeting, the FOMC made a widely-anticipated but nonetheless important change to the statement, dropping a commitment to be “patient” and thereby creating the option to hike rates at any meeting after April. In the end, this proved to be a sideshow to the more interesting changes in the Fed’s Summary of Economic Projections (SEP)—the forecasts officials’ provide at every other FOMC meeting. The changes were much larger than most observers anticipated (ourselves included), and markets reacted accordingly. Based on our proprietary measure of monetary policy surprises—which focuses on relative returns across several asset classes—last Wednesday was the third largest dovish surprise in the Quantitative easing (QE) era, only bested by the original QE1 announcement and the September 2013 “no taper” decision.

A tidy way to summarize the news is to compare changes in the Fed’s projections for the funds rate to those warranted by a Taylor Rule. This is shown in Exhibit 1 (specifically, using the midpoint of the central tendency SEP forecasts and a Taylor 1999 rule with an Okun’s Law coefficient of 2.0). Fed officials cut their forecasts for the year-end funds rate by 41 basis points (bps) for 2015, by 68bps for 2016, and by 38bps for 2017. According to a standard Taylor Rule, these changes resulted from lower core inflation forecasts and a lower estimate for the non-accelerating inflation rate of unemployment (NAIRU)/structural unemployment rate. They were partly offset by downward revisions to forecasts for the actual unemployment rate (which, of course, implies a higher projected funds rate). There is also an unexplained residual which was positive (hawkish) for 2015 but negative (dovish) for 2016 and 2017.

Exhibit 1. Changes to the Summary of Economic Projections (SEP)

FOMC Meeting Fed Meeting

At face value then, revisions to the SEP funds rate forecasts (the “dot plot”) were roughly consistent with what a Taylor Rule would have implied. Yes, there is an unexplained dovish residual for 2016 and 2017, but these are not very large numbers (they could be close to zero using a different policy rule, for example). Thus, in our view, it’s fair to say that the funds rate revisions were a roughly “data dependent” outcome given the Fed’s revisions to its economic forecasts.

Nonetheless, the meeting was surprising for two reasons. First, the revision to the NAIRU/structural unemployment rate was larger than we had expected. As the table above shows, under a standard Taylor Rule this change alone accounts for 50bps of the downward revision to the funds rate path. Without it the revisions would have been negligible.

We suspect that Fed officials are taking into account recent news on wage inflation in their assessment of labor market slack. Measures of the structural unemployment rate based on labor market data alone (i.e. the Beveridge Curve/Vacancy Supply Curve relationship) have been trending down recently, but not as steeply at the revision in the SEP (Exhibit 2). However, because of the limited uptick in wage and price inflation so far, measures of the NAIRU/structural unemployment rate derived from the relationship between inflation and labor market slack point to lower unemployment thresholds. With the latest SEP revision, Fed officials shifted in this direction. In our view, this is also consistent with data dependence, broadly defined—even if the lumpy revision at one meeting was unexpected.

Exhibit 2: Estimates of the NAIRU/structural unemployment rate (%)

FOMC Meeting Fed Meeting

The second surprise was the generally cautious comments about both growth and inflation at the meeting. While it’s factually correct that growth “has moderated somewhat” since January, we expected the statement to strike a more balanced tone given solid employment reports in recent months (Yellen did attempt this at the press conference, saying “we do see considerable underlying strength in the U.S. economy”). Similarly, the press conference materials and Yellen’s Q&A sounded marginally more concerned about market-based measures of inflation expectations. At the last meeting Yellen said that the committee was “mindful” of these developments but thought they were “likely to prove transitory.” Yesterday the statement said the committee was “attentive” (a slightly stronger word than mindful?), and no longer said the low level would prove transitory. Yellen then added in her Q&A: “And market-based measures of inflation compensation have fallen. They are low. If they were to move up over time, that would probably serve to increase my confidence [that inflation will return to target].”

Thus, the meeting as a whole was more dovish than expected. It shows a committee that’s a bit more cautious about the outlook and, ironically, willing to be more patient before reacting to the decline in the unemployment rate. That being said, we do not view last week’s meeting as a game changer for the Fed outlook more broadly. We continue to expect the FOMC to hike rates in September, with balanced risks around that meeting, and the pace of rate hikes thereafter should be faster than markets are currently pricing. In this environment, the U.S. dollar can continue to outperform and U.S. rates underperform, even with a Fed moving a bit more gingerly towards liftoff.