On Lowering the Unemployment Target, by Tim Duy: There is much buzz over the possibility the Federal Reserve will lower the unemployment rate threshold at an upcoming FOMC meeting. See, for example, a nice summary by Jon Hilsenrath at the Wall Street Journal. To be sure, such a change was already a part of the policy discussion. From the minutes of the July FOMC meeting:
Finally, the potential for clarifying or strengthening the Committee's forward guidance for the federal funds rate was discussed. In general, there was support for maintaining the current numerical thresholds in the forward guidance. A few participants expressed concern that a decision to lower the unemployment threshold could potentially lead the public to view the unemployment threshold as a policy variable that could not only be moved down but also up, thereby calling into question the credibility of the thresholds and undermining their effectiveness. Nonetheless, several participants were willing to contemplate lowering the unemployment threshold if additional accommodation were to become necessary or if the Committee wanted to adjust the mix of policy tools used to provide the appropriate level of accommodation.
To a certain extent, the issue of thresholds has taken on a new urgency as a result of the tapering debate. The Fed's excellent adventure with tapering this summer indicated that they do not fully understand the transmission mechanisms of large scale asset purchases (see Federal Reserve Governor Jeremy Stein for more). That uncertainty is another reason to wind down the program, something I think the Fed has a bias toward. But the economy has been giving little room for the FOMC to maneuver. They do not want to withdraw accommodation at this point, only to limit additional accommodation. The tightening of financial conditions this summer, however, suggested that tapering is tightening.
Consequently, the FOMC, now uncertain of the impact of ending asset purchases, wants an alternative tool for providing accommodation if needed. Enter forward guidance. The discussion was furthered by new staff studies, particularly English, Lopez-Salido, and Tetlow (2013) that illustrates the effectiveness of threshold-based forward guidance. As summarized by Hilsenrath:
The research paper—written by William English, the head of the Fed's monetary-affairs division and two other authors—argues the Fed's unemployment threshold for rate increases would be more effective if it were lower than 6.5%, possibly as low as 5.5%. In effect that would mean waiting until the job market got much better before raising rates.
The research winds, it would appear, are blowing in the direction of lowering the unemployment threshold, thereby committing to a longer period of lower rates. Not that this change will be instantaneous, or even certain. There will be resistance within the FOMC to giving up the wide degree of policy discretion under the current policy. And some believe the change will have limited impact:
San Francisco Fed President John Williams expressed skepticism Tuesday about changing the threshold. "I'm not sure in this circumstance that changing the language from 6.5 to a lower number would actually tell people on its own anything different than we're saying now," he said.
What the FOMC is saying currently is summarized by New York Federal Reserve President William Dudley:
We have established a threshold of 6.5 percent for the unemployment rate as long as we do not expect inflation to exceed 2 ½ percent at a one-to-two year horizon and inflation expectations remain well-anchored. It is likely to take a considerable amount of time to reach the 6.5 percent unemployment rate threshold. Moreover, because the 6.5 percent unemployment rate is a threshold, and not a trigger, depending on the economic circumstances, we might wait a long time after we breach the threshold before we begin to raise our federal funds rate target.
The basic story is that the Fed has already made clear that 6.5% is a threshold, not a trigger, and policymakers have indicated that they see it likely that rates will remain near zero well after 6.5% is reached. Indeed, the lower unemployment threshold has little impact on the actual timing of lift-off from the zero-bound. Table 2 of the English et al. paper reports the median date of first tightening from simulations of their model; the difference between the 6.5% and 5.5% thresholds (with the inflation threshold at 2.5%) is only one quarter (2015:3 to 2015:4). Williams can thus argue that there is little practical policy difference between the two thresholds - interest rates are likely to remain low for a very long time in either scenario.
That said, the 5.5% threshold is associated a more rapid unemployment decline and lower policymaker losses (although the share of draws for which welfare improves is slightly lower than the 6.5% scenario). Given incoming Chair Janet Yellen's interest in the costs of high unemployment, and her interest in optimal control policies of the type explored in this research, the opportunity for improvement in outcomes must be tempting. But Yellen herself has warned:
While optimal control exercises can be informative, such analyses hinge on the selection of a specific macroeconomic model as well as a set of simplifying assumptions that may be quite unrealistic. I therefore consider it imprudent to place too much weight on the policy prescriptions obtained from these methods, so I simultaneously consider other approaches for gauging the appropriate stance of monetary policy.
In short, policymakers increasingly believe they have an effective tool on hand if the process of ending asset purchases results in a undesirable tightening of financial conditions. But while the Fed is moving closer to lowering the employment threshold, this is not necessarily a certain offset for ending asset purchases. It may come into play only if the process of ending asset purchases results in a undesirable tightening of financial conditions. But it may also come into play prior to tapering. Indeed, the existing threshold is looking a little bit silly if the Fed does not start tapering soon. It is hard to see that 6.5% has any meaning whatsoever if the Fed is still purchasing assets at a rate of $85 billion a month at that point. So if the Fed has not tapered substantially as 6.5% comes into view, I would anticipate changing the threshold regardless of the forward guidance implications.
There seem to be two more under-appreciated implications from the English et al. paper. The first is not only that threshold based forward guidance works, but that it also lowers the probability of returning to the zero bound within four quarters relative to a Taylor rule. This result offers hope to those of us that worry about permanent ZIRP. The second implication is that, looking down the road, presumably threshold-based forward guidance is also a tool to help manage expectations when (hopefully) activity is stronger and inflation is a real concern. In other words, not only can it be used to replace asset purchases, but also maybe asset sales.
Bottom Line: Policymakers would like to normalize policy by moving away from asset purchases to interest rates. Emphasizing forward guidance is part of that process. Incoming research suggests not only that threshold based forward guidance is effective, but has room to be even more effective. That should be a comfort to policymakers who worry that ending asset purchases will excessively tighten financial conditions; they have a tool to change the mix of policy while leaving the level of accommodation unchanged. Whether they use it or not is another question. There has clearly been some discomfort among policymakers regarding changing the unemployment threshold. This suggests it would not necessarily be an immediate replacement for ending asset purchases. That said, it is difficult to see how the current threshold is meaningful at all if the Fed is still purchasing assets when the threshold is breached. Indeed, the current low level of unemployment relative to the threshold, combined with clear indications that the Fed has no intention of raising rates anytime soon, argues by itself that a change in the thresholds is a likely scenario in the months ahead.