Taylor Rule

Summary added on Sep 3, 2016

The Taylor Rule and related policy rules are not “right” even according to their advocates.  Rather, their main advantage is that they incorporate stabilizing feedback loops that in normal circumstances make them robust to being wrong.  Crucially, we are not now in a “normal” environment. Rather, for the past eight years where have been in the vicinity of the zero bound. In this environment, being “right” is much more important. Accordingly, the Fed cannot follow a policy rule, like Taylor, and is not following such a rule.  They have not been sufficiently plain about this, but knowing this helps explain their behavior.


The Fed is often criticized for not following a policy rule of the sort John Taylor proposed in 1993 – or even some of the more dovish analogs that were prominent earlier this decade. As Taylor himself mentioned in a Brookings debate with NY Fed President Bill Dudley back in October, “no one knows what you are doing.” That was not meant as a compliment.

My view is that the Fed’s approach to rates policy is indeed hard to predict, but that this is a function of economic circumstance rather than fecklessness among the leadership.  If I were to criticize the Fed leadership on this point it would be mostly for understating their own case and implying that the argument against a rules-based approach is other than what it actually is.

When assessing whether the Fed is clear about what it is doing, it is probably important to recognize that clarity applies both to the Fed’s objective and the specific path of the policy instruments they will use to hit that objective.  The Fed is actually quite clear on the former and unavoidably unclear on the latter.

The Fed has long been clear that its objective is to deliver an average inflation rate of about 2% over time and to minimize deviations of employment and inflation from their natural and target levels respectively.  To eliminate any doubt, they made this objective explicit in their Statement of Long Run Goals in January 2012.

I think the Fed might be criticized slightly for not being clear about how much of an employment overshoot (or unemployment undershoot) they might tolerate to reduce the period over which they are missing to the downside on inflation. But in the longer sweep of Fed history, this would seem to be dwelling in minutia, at least as regards the objectives. Until 20 years ago, it was not taken as self evident that the Fed should even have clear objectives. Their power was meant to reside partly in the mystery.

The question of how the Fed will meets its objective is definitely secondary, in terms of the central bank’s contribution to economic welfare. A business manager, for example, would like to have some sense of how nominal demand growth will run in the next few years. Unlike me, he doesn’t enjoy arguing about the portfolio balance channel of QE or in how to estimate the equilibrium real interest rate!  This distinction is often overlooked.

Back in October, John Taylor reiterated his case for basing policy on a Tayor Rule, although one nested in a broader rule-based framework that would allow the Fed to deliver stimulus in an environment of – for example – the zero bound on rates.  I thought Dudley’s take on this – which he actually delivered ahead of Taylor – was telling for missing the main point.

Dudley claimed that the Taylor Rule is too reductionist, fails to recognize that the neutral real rate can move, misses the role of financial market developments, and does not take account of either forecasts or the fact that the Fed cannot follow Taylor when it and rules like it prescribe additional stimulus with the funds rate already at zero. Those criticisms may or may not be true, but they would seem to have little to do with why the Fed has recently followed maximum discretion and has seemed to be flying blind. (The last criticism comes close, but does not quite hit the mark, because it is about delivering stimulus rather than about calibrating new restraint, as I will describe below.)

To understand the issue here, I think it is useful to focus on what is meant to be the key advantage of the Taylor Rule, which in the interest of brevity I will take as a stand in for the entire class of rule-based strategies.

Crucially, the Taylor Rule does not identify the level of the funds rate appropriate to the current position of the business cycle. It would be a bizarre coincidence, as Dudley emphasizes, if the funds rate exactly appropriate to today could be described as a linear combination of the unemployment rate (or output gap) and the inflation rate.  The world is much more complex than that, although this point too is often overlooked.  I think Taylor himself contributes to the mischief when he implies that the difference between the Taylor-prescribed rate and the actual funds rate is a measure of policy misalignment. It most certainly is not. And more to the point, the intellectual case for the Taylor Rule does not assume it is!

No, the key feature of the Taylor Rule is that it incorporates stabilizing feedback signals that allow it to perform well in most economic environments (at least as simulated by econometric studies) despite its failure to identify the appropriate level of the fund rate in real time (which would be impossible).

For example, if the Taylor Rule were to prescribe a funds rate of 4%, and if that turned out to be too restrictive, then the following sequence of largely-benign events would arise.   The economy would slow in an undesired way and either the unemployment rate would rise or the inflation rate would fall. This would be recognized by the Taylor Rule as a signal that easier policy was required. And according to econometric simulations that endorse the Rule, those signals would arrive sufficiently quickly to allow the Rule to deliver satisfactory macro economic conditions over time, even though it would be absurd to think that the Taylor Rule “knows” what the appropriate fed funds rate is in real time.

Let me put this starkly in the interest of clarity. The Taylor Rule is not right. Rather, it is robust to being wrong.  That is really central to understanding the Rule’s virtue and the virtue of rules that are close analogs to it.

However, none of that applies in the current environment, which is why following the Taylor Rule or allied approaches is ruled out.  Right now, the Fed needs to make a judgment call on how its rates policy path will affect the economic outlook and cannot rely on the central feature of the Taylor Rule, which is that mistakes are easily corrected. This is admittedly a sticky situation to be in. I do not envy them.

If the Fed raises interest rates (too quickly) now and the economy starts to cycle down and away from the Fed’s objective, then the Taylor principle will advise that policy reverse course and ease.  But in that environment, the Fed will not be able to reverse course and ease (at least not much) because the starting position for the funds rate is already near zero and the unconventional tools are either of dubious efficacy, ruled out by politics or both.

So the Fed is not following a Taylor-like rule. And yes, that does mean that they seem to be applying a lot of discretion in their choice of how to pursue their objective. I share Taylor’s view that it would be better to live in a world where a more rule-oriented approach could be applied.  But we do not currently live in such a world; it would risk disaster to pretend otherwise; and that explains why the Fed is not pretending.  It actually has very little to do with the quibbles Dudley raised against Taylor in their October discussion, at least in my view.

I would like to conclude on this by reiterating that this situation is indeed unfortunate and makes the Fed’s behavior very difficult to forecast. In an environment in which the Taylor Rule could be applied, we would normally think that a falling unemployment rate implies a steadily rising appropriate level of the real funds rate.  But such reductionism is actually false, and dangerously so in the current environment. We could easily imagine a situation where the unemployment rate were to fall to a very low level, and yet the level of the real funds rate required to generate a recession might still be quite low. In other words, today, the intuition of the Taylor Rule might be all wrong, not only as a guide for policy makers, but as a guide to those of us trying to predict them or the economy itself.

So, yeah, it is going to look quite confusing for a while. For that we can blame the state of the world and maybe even past Feds.  But this is not really the current Fed’s doing. They are simply reacting to the world we actually live in.