The FOMC’s policy rule, as inferred from their rate decisions and projections, appears to have changed dramatically from the previously dominant balanced-approach rule, set out by Yellen in a 2012 speech and still featured today in the Monetary Policy Report.1 We see that rule as consistent with the rapid decline in rates and with the unconventional policies pursued after the funds rate reached the effective lower bound.
However, once the unemployment rate neared and then fell below the NAIRU, in about late 2015, the FOMC’s rate decisions and projections suggest that the FOMC’s response to changes in the unemployment rate became dramatically less aggressive. Indeed, the FOMC appears to be almost unresponsive to changes in the unemployment gap. I consider two policy rules—both derived by making simple adjustments to the parameters in the balanced-approach rule—whose prescriptions seem to align better with the FOMC’s apparent reaction function.
1 See Janet L. Yellen’s (2012), “The Economic Outlook and Monetary Policy,” speech delivered at the Money Marketeers of New York University, New York, New York, April 11.
The Policy Rule After the Great Recession
That aggressiveness is displayed in Figure 1 by the rapidity of the cuts in the fund’s rate as it fell to the zero lower bound and in the prescriptions of a negative real funds rate, troughing below -6%, from roughly 2009 through 2013. The gap between the zero lower bound and the prescribed fund’s rate was, in principle, to be filled by the unconventional policies which followed—including balance sheet policies, forward guidance, and holding the fund’s rate at zero longer than the rule prescribed.
Headwinds and the Short-run and Long-run r-star
Yellen revised her policy rule in 2015 to reflect the unique headwinds in the post-Great-Recession period.4 The headwinds called for a lower funds rate than in normal times. She lowered her estimate of r-star to near zero, the estimate based on the Laubach and Williams model, and referred to that as the “short-run” neutral rate. The prescription from the rule, using a lower short-run r-star, was for a near-zero nominal funds rate, supporting the idea that it remained appropriate for the FOMC to keep the fund’s rate near zero even though the balanced-approach rule, using participants’ median estimate of r-star, prescribed tighter policy. This is shown in Figure 1. Yellen expected r-star to converge to this estimate of its long-run value as the headwinds dissipated, which would make the balanced-approach rule relevant again.5
2In that speech, Yellen used a coefficient on the unemployment gap of 2.3, which was an Okun’s Law translation from a rule using a coefficient of 1.0 on the output gap. However, we use 2.0 because that’s what she used in a subsequent speech and that’s also the coefficient used in the Monetary Policy Report today.
3 Participants report their estimates of the nominal fund’s rate and the unemployment rate in the long run. The median estimate of the nominal neutral fund’s rate is converted to a corresponding real estimate by subtracting 2, which is the FOMC’s inflation objective. The median estimate of the unemployment rate, in the long run, is equivalent to the NAIRU.
4 See Janet L. Yellen (2015), “Normalizing Monetary Policy: Prospects and Perspectives,” speech delivered at the “The New Normal Monetary Policy,” a research conference sponsored by the Federal Reserve Bank of San Francisco, San Francisco, California, March 27. 5 Today the discussion between short-run and long-run values for r-star has mostly been dropped, though it occasionally resurfaces.
The Policy Rule Today
Yellen’s distinction between the levels of r-star in the short-run and long-run appears to have disappeared after the FOMC began to raise rates in late 2015. It was certainly gone by 2017 and 2018, the years we focus on here. Note that the balanced-approach rule would have called for a much earlier exit from near-zero interest rates and a rise in the fund’s rate to 4% beginning in 2018. That’s sharply above the target range today, and well above the path of the fund’s rate through 2021 in participants’ March 2019 rate projections. Indeed, that rule prescribes a move well into restrictive territory, whereas the March projections show a funds rate remaining near the neutral rate. So, if they are not following that rule, the question is what rule they are following today. Is it dramatically different from the balanced-approach rule and, if so, why?
A Rule Consistent with the Flat Phillips Curve
The response of the FOMC to shocks should depend on the structure of the economy, and it should therefore evolve as that structure changes.6 Perhaps the most important change in the structure of the economy with respect to the conduct of monetary policy—other than the decline in the real neutral rate—is the flattening of the Phillips curve. The flatter the Phillips curve, the smaller the overshoot of the inflation objective that results from an undershooting of the NAIRU. That, in turn, warrants a more-tempered response to declines in the unemployment rate below the NAIRU—at least in current circumstances. As Clarida put it in his speech last week: “A flatter Phillips curve is, in a sense, a proverbial double-edged sword. It permits the Federal Reserve to support employment more aggressively during downturns—as was the case during and after the Great Recession—because a sustained inflation breakout is less likely when the Phillips curve is flatter. However, a flatter Phillips curve also increases the cost, in terms of economic output, of reversing unwelcome increases in longer-run inflation expectations.” But in current circumstances, things are not so symmetric. In particular, inflation has languished below 2%, and inflation expectations are too low if anything. As Clarida noted, well-anchored inflation expectations are particularly important with a flat Phillips curve. A Near-Zero Unemployment Gap
In 2012, the NAIRU was estimated to be 5½%, according to the median FOMC projection of the longer-run unemployment rate. That estimate has since fallen to only 4.3%. But some on the Committee see the NAIRU as still lower. For some, there is little if any unemployment gap today. A policy rule with a zero or near-zero unemployment gap is observationally equivalent to a rule with a near-zero coefficient on the unemployment gap.
Indeed, Clarida recently said that “The range of plausible estimates [of the NAIRU] likely extends at least as low as the current level of the unemployment rate,” which is 3.8%. As an example, he noted that the average of the 10 lowest estimates from Blue Chip survey respondents was 3.9%, compared to 4.7% for the highest 10.7 However, no participant at the March meeting had an estimate of the NAIRU as low as 3.9%, and 10 had estimates between 4.2% and 4.5%.
There is a powerful connection between the assumptions in the two rules. The flatter the Phillips curve, the harder it is to discern that the unemployment rate has diverged from the NAIRU. That makes estimating the NAIRU more uncertain. Indeed, given the flat Phillips curve, an unemployment rate within a range of ½ pp in either direction—and perhaps more—from the NAIRU is indistinguishable from being at the point estimate of
6 The dominant change in the structure of the economy with respect to monetary policy is the decline in real rates that will likely result in the FOMC being constrained by the zero lower bound more frequently. This is, in my view, the primary reason for the review of strategy and communications currently underway.
7 See Richard Clarida (2019), “The Federal Reserve’s Review of Its Monetary Policy Strategy, Tools, and Communication Practices,” speech delivered at the “Fed Listens: Distributional Consequences of the Cycle and Monetary Policy,” a conference hosted by the Opportunity and Inclusive Growth Institute, Federal Reserve Bank of Minneapolis, Minneapolis, Minnesota, April 9.
the NAIRU in terms of the inflation outcome. Indeed, that is what we are seeing today. It is as if we are at the NAIRU, and maybe we are.
Reverse Engineering the Policy Rule Today
In a December 2017 commentary (link), I tried to reverse engineer the implied policy rule based on participants’ median rate projections reported in the September 2017 SEP and concluded that they implied that the unemployment gap was zero or that the coefficient on the unemployment gap was closer to zero than to the 2.0 in the balanced-approach rule or even the 1.0 in Taylor’s original rule. Indeed, Yellen’s balanced-approach rule prescribes a roughly 4% funds rate now through 2021, whereas the median projection has the fund’s rate staying near 2½%. We update that analysis here by reverse-engineering the policy rule implied from the median rate projections in March 2019.
In Figure 2, we show the actual funds rate as well as participants’ median projection for the fund’s rate from the March 2019 SEP through 2021; the prescription from the balanced-approach rule using the SEP r-star; and prescriptions from the two alternative policy rules discussed above.
One of the alternative rules assumes a nearly flat Phillips curve. In this case, we assume the coefficient on the unemployment gap in the rule is zero. In the second case, policymakers perceive that the economy has been at full employment during this period, not beyond it; the unemployment gap is therefore zero. In either case, the term with the unemployment gap drops out of the policy rule and the two rules are therefore observationally equivalent. As a result, we need only one line in Figure 2 for these rules because the prescriptions are the same.
We can see that the prescription from the alternative rules is in line with that from the balanced-approach rule when the unemployment rate nears participants’ median estimate of the NAIRU in late 2015. But, once the unemployment rate fell below the estimated NAIRU, and continuing in the projections through 2021, the prescription from the alternative rules is more consistent with FOMC policy, both actual rate decisions to this point and the projected rate path. That is why we refer to the alternative rules as the FOMC’s policy rule today.
The FOMC’s Policy Rule in the Future
We cannot judge how long monetary policy will continue to align with these alternative rules. Given the flat Phillips curve, we expect the FOMC’s reaction function will no longer call for an aggressive response to deviations from the NAIRU, at least when the unemployment rate is below the NAIRU. If inflation picks up earlier than the FOMC projects, and faster than implied by the Phillips curve today, we expect the FOMC would revise up its estimate of the NAIRU and/or increase the weight in the policy rule on the unemployment gap. There may also be some unemployment rate threshold at which the Committee wants to take out insurance against the rising risk that inflation will rise more than acceptable, perhaps because the Phillips curve may exhibit nonlinearity and become steeper. In any case, one feature of the FOMC’s reaction function that is more clear is that, given the proximity of the zero lower bound, the FOMC will be more preemptive and more aggressive in lowering the fund’s rate as the economy slows to a below-trend rate and as recession risk is rising.