What do you mean by policy lags?

Economics textbooks describe at least three types of policy lags—the recognition lag, the implementation lag and the impact lag. But even that may not be enough, as the impact lag is a rather ambiguous concept, and might include two separate lags:

1. There is the lag between when a central bank takes concrete steps to ease or tighten policy, and the time when the policy stance actually changes.
2. There is the lag between the time when policy stance actually changes, and the time when aggregate demand responds to that policy change.

The first lag could be five days, five weeks, five years, or even five centuries.  Consider the following hypothetical:

Assume the central bank has fallen “behind the curve”, as inflation has increased because interest rates were held below their natural rate.  The central bank raises interest rates by 1% each year, but the natural rate rises by 1.1% each year (mostly due to the Fisher effect.)  After 100 years, nominal rates have risen by 100% (i.e., 10,000 basis points), and yet the central bank has fallen ever further behind the curve, as the natural rate has risen even faster.  (Here you can think of real interest rates as having fallen.)  We are now experiencing triple digit inflation.  Then a frustrated central bank suddenly raises interest rates dramatically, the economy falls into recession, and inflation plunges much lower.

How should we interpret that thought experiment? Is that a case of central bank tightening causing inflation to fall after a 100-year lag?  Or was their no real tightening at all, at least until the sudden dramatic rise in rates after 100 years had passed?  I lean toward the latter view, but I’m not certain my definition is universally accepted.

Here it’s useful to get beyond quibbles over terminology and focus on where there is broad agreement.  Most economists agree that rising interest rates do not represent tight money if the natural interest rate is rising even faster.  When the Fed began raising rates in early 2022, Jay Powell referred to the need to first get rates up to “neutral”, which is roughly what economists mean by the natural, or equilibrium, rate of interest.  The implicit assumption was that as long as rates are below neutral, the net effect of monetary policy is still expansionary.

Nonetheless, it seems to me that most pundits (wrongly) talk about the impact policy lag in an inclusive way, including both the lag between concrete policy steps and an effective change in policy, as well as the lag between an effective change in policy and a change in aggregate demand.  But how can this be reconciled with the frequent claims that the impact lag is something like 6 to 18 months. From these estimates, it’s pretty clear that the term “impact lag” should not refer to the lag between concrete steps and an effective change in policy.  After all, in my hypothetical example above that lag was 100 years!  And it could be any figure at all; it entirely depends on how the policy rate is adjusted relative to the natural interest rate. 

The lag between concrete actions and changes in the policy stance is not a parameter “out there” in the world, waiting to be estimated by econometricians; it is a figure that entirely depends on the skill of monetary policymakers.  So when economists speak of a 6 to 18 month policy lag, they must be talking about the lag between an effective change in the stance of policy and the growth rate of aggregate demand.  It would be utter nonsense to speak of a 6 to 18 month lag between concrete policy steps and an effective change in the stance of policy.

But effective policy changes are hard to observe, as the natural rate of interest is not a variable that can be directly measured.  And in the rare cases where the change in policy is so dramatic that it can be easily identified, such as the dollar depreciation that began in April 1933, the effects seem almost instantaneous.  Nominal GNP rose by 23.7% between 1933:Q1 and 1933:Q3.  And that’s the actual increase in just 6 months, not the annualized rate of change!

So why do economists estimate relatively long impact lags?  I suspect it’s partly the identification problem.  These economists might be (wrongly) including some of the lag between concrete actions by the central bank and their impact of the stance of monetary policy.  Thus some people seem to assume that the Fed’s been tightening policy all year, and that we haven’t seen much slowdown in demand due to policy lags.  In fact, interest rates probably remained well below neutral for much of the year, and the Fed only began tightening recently.  Perhaps they still have not begun to tighten policy.

Because there’s no general agreement about how to define tight money, let’s focus on a more clearly defined issue—policy counterfactuals.  I believe that if the Fed had been told a year ago how fast NGDP would grow over the subsequent 12 months, they would have preferred a tighter policy.  A year ago, NGDP had already moved above the pre-Covid trend line.  Even if you reject the “level targeting” argument for going back to the previous trend line, surely it was not wise to go further and further above the previous trend line.  At a minimum, the Fed should have tried to get NGDP growth to level off at no more than roughly 4% or so.  Instead, it rose by 9% over the past 4 quarters, and 2022:Q4 looks like another hot one.

To prevent further overshoots, the Fed needed to quickly get interest rates up to a neutral level.  Instead, they raised rates at a very slow pace, which meant that policy remained effectively expansionary through most of 2022.  The problem is getting worse as each month goes by.  We aren’t making progress, just the opposite.

There’s a great deal of debate about the policy mistakes that led to the initial inflation overshoot in 2021.  Was it easy money, fiscal stimulus, supply shocks, or all of the above?  I see much less debate about the failure of Fed policy to effectively address the issue once it was clear that aggregate demand was rising much too rapidly.  Why not immediately raise rates to at least neutral?  Perhaps the Fed feared overshooting toward a highly contractionary policy, and thus triggering a recession.

These mistakes are much more likely to occur when central banks lack a clear policy regime, a clear target path for a nominal aggregate such as NGDP or core inflation.  Without that sort of clear policy target, it’s very difficult to know what sort of interest rate setting is appropriate.  Fed funds futures provide little guidance, as they merely predict future interest rates, not future appropriate interest rates.  I don’t know what interest rate would have been appropriate in 2022, because if NGDP growth had been only 4%, then the “natural” interest rate would have been far lower.  I’m actually not sure the Fed needed higher interest rates throughout all of 2022, I am sure they needed higher interest rates relative to the natural rate.

This is why I have trouble answering commenter queries about what monetary policy would have been appropriate.  They want me to describe an alternative path of interest rates, whereas I can only describe an alternative monetary regime—NGDP level targeting.