The science of monetary policy has not progressed to the point where we can specify the optimal rule for setting monetary policy. The reason is that optimal rules, that is, those that maximize economic welfare, are highly dependent on the particular model from which they are derived, and there is no broad-based consensus for the right model. More relevant is the finding that the optimal rule for one model can produce very bad outcomes in another model. In addition, optimal rules can often be quite complex, thus making them difficult to implement and to communicate to the public. In other words, they may not be very transparent.

However, these limitations to implementing optimal policy rules should not deter us from efforts to adopt a more systematic rule-like approach to the conduct of policy. There has been a great deal of progress made in identifying simple rules that appear to perform well in a variety of models and environments. Such robust rules can form a basis for developing more systematic, rule-like policymaking.

One important and desirable characteristic of a systematic and rule-like approach to policy relates to communication. In particular, it is an approach that is easily communicated to the public and thus greatly improves the transparency and predictability of monetary policy, which reduces surprises. The public and markets are more informed about the course of monetary policy because they understand how policymakers are likely to react to changing economic circumstances. Equally important in my view is that greater clarity about the policymakers' reaction function strengthens accountability and thus can serve to preserve the central bank's independence.

Given model uncertainty and data measurement problems, there are, of course, limitations to the use of a simple rule. The rule is basically intended to work well on average, but central banks look at many variables in determining policy. There inevitably will be times when economic developments fall outside the scope of our models and warrant unusual monetary policy action. Events such as 9/11, the Asian financial crisis, the collapse of Lehman Brothers, and the 1987 stock market crash may require departures from a simple rule. Having articulated a rule guiding policymaking in normal times, however, policymakers will be expected to explain the departures from the rule in these unusual circumstances. With a rule as a baseline, departures can be quantified and inform us how excessively tight or easy policy might be relative to normal. If the events are temporary, policymakers will have to explain how and when policy is likely to return to normal. Thus, a simple rule provides a valuable benchmark for assessing the appropriate stance of policy. That makes it a useful tool to enhance effective communication and transparency.

One of the most important ways to support credibility and thus the effectiveness of forward guidance is to practice it as part of a systematic policy framework. I believe that indicating how the evolution of key economic variables systematically shapes current and future policy decisions is critical to such a policy framework. Indeed, a commitment to a policy framework that is systematic and rule-like provides the foundation for establishing expectations for the future path of policy and thus forward guidance...

Systematic policies that provide important information about the policymakers reaction function combined with other information, such as the policymakers economic forecasts, can sharpen forward guidance in ways that reduce policy uncertainty and enhance economic performance. Thus, well-designed communications are valuable, and behaving systematically has the added advantage of making those communications easier for the public to understand.

HENSARLING: I will say this if I can, Madam Chair, there's one thing that the Fed says. It's another that markets may hear.
My time is running out. I want to cover a little other ground, as well dealing with a rules based monetary policy.
I think if I have read some of your statements properly -- and I don't want to put words in your mouth -- but that you consider times five years after the financial crisis still extraordinary, and it is not necessarily an appropriate time for a rules-based approach?
Is that a fair assessment of your views?
YELLEN: So I have always been in favor of a predictable monetary policy that responds in a systematic way to shifts in economic variables...
HENSARLING: Well, in fact, earlier in your career, with respect -- in reference to the Taylor rule you said it is, quote, "What sensible central banks do," unquote.
So that begs the question today, using your words, are you a sensible central banker, and if not, when will you become one?
YELLEN: Congressman, I believe that I am a sensible central banker, and these are very unusual times in which monetary policy for quite a long time is not even been able to do what a rule like the Taylor Rule would have prescribed. For several years, that rule would have prescribed that the federal funds rate should be in negative territory, which is impossible.
So the conditions facing the economy are extremely unusual.
I have tried to argue and believe strongly that, while a Taylor Rule is -- or something like it -- provides a sensible approach in more normal times, like the great moderation, under current conditions when this economy has severe headwinds from the financial crisis, and has not been able to move the funds rate into the negative territory that rule would have prescribed that we need to follow a different approach. And we are attempting, through our forward guidance, to be as systematic and predictable as we can possibly be.

In my view, rather than try to maintain discretion, policymakers would achieve better economic outcomes and greater clarity by taking a systematic approach to policy. But how do we get there from here? I think we could vastly improve policy going forward by doing three things, which would begin to normalize monetary policy.

• The first step is to wind down our asset purchases by the end of the year in a gradual and predictable manner. As I said, I see little if any benefit from these purchases, and growing costs.
• The second step is for the FOMC to commit to its forward guidance on the fed funds rate path, that is, to begin treating the 6.5 percent unemployment rate and the 2.5 percent inflation rate in the guidance as triggers rather than thresholds.
• The third part of the strategy is to provide information on how our interest rate policy will evolve after the trigger is reached. A commitment to a robust policy rule, perhaps consistent with the way policy was conducted prior to the crisis, would provide needed clarity on how the Committee intends to vary its policy in response to changes in economic conditions.

These steps form part of a systematic approach to policymaking. They embody clarity and commitment. By helping the public and market participants form more accurate judgments about the future course of policy, systematic policymaking can improve the efficacy of monetary policy.

I first described the ideas [of my "liftoff plan"] about three weeks ago, in Ironwood, Michigan. It evinced mixed reactions. Some observers felt that the proposed liftoff plan was dovish, in the sense that it seemed to put a lot of weight on the employment mandate. Others argued that the plan was hawkish, in the sense that it put a lot of weight on the price stability mandate.

I think that this dispersion of views reflects a simple fact: The plan is neither hawkish nor dovish. The terms “hawkish” and “dovish” presume that the Committee faces a tension between its two mandates. But the Committee does not see any tension between its two mandates now.

Lockhart told reporters after his speech he was skeptical about a proposal made yesterday by Minneapolis Fed President Narayana Kocherlakota that the central bank should hold interest rates at around zero until unemployment drops below 5.5 percent.

“My comment on that is 5.5 percent is much closer to current measures of full employment,” Lockhart said. “I am not so sure that it will not be appropriate to begin the process of tightening before we would get to what we would consider full employment. I don’t want to rule out that could be the case.”

My main message today is that the FOMC can provide additional monetary stimulus by making this sentence more precise in the form of what I’m going to call a liftoff plan: a description of the economic conditions that would lead the Committee to contemplate the initial increase in the fed funds rate above its currently extraordinarily low level.

I will suggest the following specific contingency plan for liftoff:

As long as the FOMC satisfies its price stability mandate, it should keep the fed funds rate extraordinarily low until the unemployment rate has fallen below 5.5 percent.

The fed funds rate is a short-term interbank lending rate that is the FOMC’s usual vehicle for influencing credit conditions. I’ll be much more precise later about the meaning of the phrase “satisfies its price stability mandate.” Briefly, though, I mean that longer-term inflation expectations are stable and that the Committee’s medium-term outlook for the annual inflation rate is within a quarter of a percentage point of its target of 2 percent. The substance of this liftoff plan is that, as long as longer-term inflation expectations remain stable, the Committee will not raise the fed funds rate unless the medium-term outlook for the inflation rate exceeds a threshold value of 2 1/4 percent or the unemployment rate falls below a threshold value of 5.5 percent. Note that neither of these thresholds should be viewed as triggers—that is, once the relevant cutoffs are crossed, the Committee retains the option of either keeping the fed funds rate extraordinarily low or raising the fed funds rate.

This is a remarkable change of views.  See his comments from earlier this year.

As I noted, simple rules have the advantage of delivering good policy outcomes across a broad range of models, and are thereby relatively robust to our limited understanding of the precise working of the economy--in contrast to optimal-control policies, whose prescriptions are sensitive to the specification of the particular model used in the analysis. However, simple rules also have their shortcomings, leading them to significantly understate the case for keeping policy persistently accommodative in current circumstances.

One of these shortcomings is that the rules do not adjust for the constraints that the zero lower bound has placed on conventional monetary policy since late 2008. A second is that they do not fully take account of the protracted nature of the forces that have been restraining aggregate demand in the aftermath of the housing bust...

I have proposed that any further accommodative policies should contain a safeguard against an unreasonable increase in inflation. In my judgment, nominal income level targeting is an appropriate policy choice and has such a safeguard. But recognizing the difficult nature of that policy approach, I have a more modest proposal: I support a conditional approach, whereby the federal funds rate is not increased until the unemployment rate falls below 7 percent, at least, or if inflation rises above 3 percent over the medium-term. The economic conditionality in my 7/3 threshold policy would clarify our forward policy intentions greatly and provide a more meaningful guide on how long the federal funds rate will remain low.

I consider a highly accommodative policy stance to be appropriate in present circumstances. But considerable uncertainty surrounds the outlook, and I remain prepared to adjust my policy views in response to incoming information. In particular, further easing actions could be warranted if the recovery proceeds at a slower-than-expected pace, while a significant acceleration in the pace of recovery could call for an earlier beginning to the process of policy firming than the FOMC currently anticipates.

More broadly, these considerations help illustrate why it would be imprudent to adhere mechanistically to the prescriptions of any single policy rule. Such rules can serve as useful benchmarks for facilitating monetary policy deliberations and communications, but a dose of good judgment will always be essential as well.

I believe that the Fed should provide more information about its reaction function. The practice of using systematic rules as guides to monetary policy imposes an important discipline on policymaking and improves communication and transparency. This is because systematic rules make policy more predictable and therefore helps the public and markets make better decisions. Moreover, if policymakers choose to deviate from the guidelines, they are forced to explain why and how they anticipate returning to normal operating practices. Systematic policy also reduces the temptation to engage in discretionary policies.

I believe the Committee is still some way from agreeing on one systematic policy rule or reaction function. Such choices will involve elaborate discussions and agreement on the appropriate class of models and an agreed-upon loss function. One way to move toward more systematic policy would be to describe the variables that are important for our response function.

The standard response to {a projected 0.7% increase} in core PCE inflation would be to raise the target interest rate by a larger amount—that is, by at least 70 basis points. For example, the widely known rules associated with John Taylor of Stanford University would recommend that the response should be to raise the target interest rate by 1.5 times the increase in core inflation—that is, by 105 basis points...

[T]he Fed should adopt an explicit numerical inflation objective. Moreover, in my view, now is an opportune time to do so. The apparent strengthening of the U.S. economy suggests that, in the not-too-distant future, monetary policy will have to begin reversing course from a very accommodative policy stance. As we choreograph that exit, I believe that the Fed should do all it can to underscore its commitment to maintaining price stability.

...

Some people may argue that there is no need to articulate a numerical inflation objective because the Fed has established a strong record of maintaining low and stable inflation over the last two decades. But this is not an argument against an explicit target. It is an argument against commitment.

I want to highlight the analogous character in practical effect of traditional monetary policy using interest rates and the less familiar asset-purchase tools we have employed since the federal funds rate hit its lower bound. Though some have argued otherwise, I believe the FOMC hasoperated for at least a decade with a consistent and fairly well understood rules-based framework. It is within this framework that I think about the desirability of both LSAP3 and the inevitable exit to a less accommodative policy stance.

...

Even though I personally am not expecting an immediate need to implement an exit, I think it's fully appropriate to revisit the implementation assumptions and tools readiness. As I contemplate an exit, two basic and obvious questions come to mind—when will it be appropriate to undertake an exit, and how to implement the exit.

...Since I think passive unwinding is probably not feasible, we will have to decide when to actively implement an exit strategy. And though the answer to when to do this is clear in concept, it is not straightforward in practice. Lags in the effects of monetary policy mean that action generally needs to be taken in advance of definitive changes in the path of economic activity and prices. That is why the policy framework I am describing emphasizes a forward-looking rule-like construct to which the FOMC would simply react.