The commentary on the possibility of the Fed lowering the unemployment threshold left me a little surprised this morning. It seems that people are jumping to the conclusion that a 5.5% unemployment threshold implies that the Federal Reserve would intend to hold rates near zero until 2017. For instance, Gavyn Davies at the FT writes:
Compared to previously published simulations, the new ones in the English paper are even more dovish. They imply that the first hike in short rates should be in 2017, a year later than before. More interestingly, they experiment with various thresholds that could be used to persuade the markets that the Fed really, really will keep short rates at zero, even if the economy recovers and inflation exceeds target. They conclude that the best way of doing this may be to set an unemployment threshold at 5.5 per cent, which is 1 per cent lower than the threshold currently in place, since this would produce the best mix of inflation and unemployment in the next few years.
Edward Harrison agrees:
In the one paper by Fed economist William English (pdf here), we see an argument for lowering the threshold for rate increases to 5.5% instead of the present 6.5%. The argument here is that the situation today requires significantly more stimulus than a Taylor rule or the current thresholds for employment and interest rates imply. Moving the thresholds would mean rates would not rise until at least 2017, according to current Fed thinking about the economy.
I think this is a misinterpretation of the paper. The 2017 date results from the optimal control analysis with policy commitment. The 5.5% threshold comes from an analysis of threshold policies that attempt to mimic optimal control outcomes. Regarding the optimal control results, the authors state:
The discussion in the previous section suggests that, given the current outlook, strategies to provide additional stimulus would be consistent with achieving outcomes better aligned with the assumed long-run policy goals. Nevertheless, such “optimal” policies might be viewed as theoretical references of limited usefulness in Committee communications, because they are both complex and model-dependent, and because they do not reveal how the Committee would respond to changes in the economic outlook.
Notes that this echoes the words of incoming Federal Reserve Chair Janet Yellen:
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.
What is the policy difference between the optimal control and the threshold-based forward guidance approaches? Refer to Figure 7 of the paper:
The difference in lift-off dates between the 6.5% and 5.5% thresholds is very small, just one quarter. This is confirmed by the simulation results presented in table 2:
The expected timing of first rate hike moves forward by just one quarter; both remain in 2015, consistent with existing policymaker expectations of the expected timing of liftoff. Hence why San Francisco Federal Reserve President John Williams is a bit befuddled by the excitement:
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.
In short, if you focus on the possibility of lowering the unemployment thresholds to 5.5%, you should expect only a minor shift in the timing of the first rate hike. The reason for this is obvious - everyone already believes that, under the current circumstances, the 6.5% threshold is no longer meaningful. No one expects a rate hike when the 6.5% mark is crossed. A threshold of 5.5% is largely just a recognition of the reality of the likely policy path.
In contrast, the 2017 number falls out of the optimal control problem in which the Fed credibily commits to holding rates near zero through that date. That is a different policy than the threshold based guidance currently in play. And I would say that persistent concerns about financial stability make it difficult for the Fed to credibly commit to such a period of low rates, even if policymakers wanted to make such a commitment. Hence the exercise with threshold-based guidance to begin with - it is intended to capture as many of the gains of the optimal control approach as possible assuming the optimal control approach is not a realistic policy option.
Bottom Line: Be wary of assuming that a change in the unemployment threshold to 5.5% implies the Federal Reserve intends to keep rates near zero into 2017. The results of the English et al. (2013) paper suggests a much smaller change in expectations for the timing of the first rate hike from such a policy shift.