What is the US Federal Reserve Aiming At?

With the confirmation hearings for the two top posts at the Federal Reserve imminent, Peter Doyle argues that this question needs to be posed and answered now.

Post Date
10 January, 2022
Reading Time
6 min read
US Federal Reserve symbol

The pandemic has given the new US Federal Reserve monetary framework, introduced in late 2020, a baptism of fire.

That framework, denoted Average Inflation Targeting (AIT), aims to secure inflation of 2 percent on average over a number of years and is defined on the PCE measure of consumer prices.

The averaging procedure has two effects:

  • to boost monetary firepower when it is most needed—by pre-committing the Fed to balance any
    inflation target undershoots with later overshoots. This permanently combats risk that low-flation in recession at the effective lower bound becomes entrenched by depressing nominal wage growth;
  • to penalize officials who, when in doubt (as they always are), presume to aim below rather than at the 2 percent inflation target. Given persistent inflation undershoots, that behaviour is a coherent explanation of post GFC output as opposed a second coming of some mythical “secular stagnation”.

But after sustained inflation target undershoots in the last decade, Covid struck from late 2019, raising inflation well above target and generating high uncertainty about the outlook for output (see chart below).

So a key question is whether the Federal Reserve, in accord with AIT, should now declare that it will aim for a period of below-target inflation to correct for pandemic overshoots.

A challenge in answering is that the Fed has not specified how long its averaging window is.

That is no small matter, as can be inferred by illustrating with four, six, and ten year averaging windows centred on the latest inflation observation, first for November 2021, and second—assuming PCE inflation in the next six months at 5.7 percent—for May 2022.

On this basis, given outturns in the first half (H1) of each of these windows, the targets for each of the second halves (H2) of each averaging window are as follows:

Average PCE Target in percent

Averaging window centred on
November 2021 May 2022
H1 outturn H2 target H1 outturn H2 target
Four-year window 2.6 1.4 3.7 0.3
Six-year window 2.3 1.7 3.0 1.0
Ten-year window 2.2 1.8 2.6 1.4

May 2022 assumes that PCE inflation is 5.7 percent until May 2022

Though these H2 target numbers are a bracing wake-up call, even they do not convey the full implications. Given that PCE inflation is now running over 6 percent, it would have to approach zero if not go negative in the latter part of 2023 in order to meet the H2 average target of 1.4 percent for the four-year window centred on November 2021. Similarly, though less dramatically, for the longer windows for November 2021, and much more dramatically for the window centred on May 2022 assuming inflation from now to May of 5 1/2 percent.

All this cuts across the established trench lines between acrimonious economists:

  • Those who damned AIT at its birth—largely because they approved of policymakers aiming-below target— most of whom have also been in the vanguard of pandemic inflation alarums now find themselves suddenly pleased that AIT leans hard and possibly very hard into policy tightening.
  • Those who cheered framework reform as they not only opposed aiming below target but also
    favoured a central target of 3 or 4 percent are dismayed that rather than opportunistically using the
    pandemic to raise the central target, AIT is set to drive inflation down too far, possibly below zero.

So in pandemic, albeit for very different reasons, no-one ends up happy with AIT as is.

In this context, a minimal framework adjustment would be to adopt formally a long averaging window – perhaps ten years. But as noted, that implies an average inflation target for the next five years of 1.4 percent, implying even lower inflation during the period to offset the high numbers now. So while that adjustment in the spirit of AIT would reinforce efforts to stop current inflation from feeding into wages and would do so at less risk of overkill than shorter windows, the clear risk of overkill remains.

That is because AIT, as is, reflects a fundamental conceptual mistake which the pandemic has exposed.

In particular, the original (?) proposal for adoption of such windows conceived them as error correction not averaging procedures, motivated to boost monetary firepower at the lower bound and discipline systematic off target aiming by the FOMC. Given those twin motivations, such target error corrections should apply only to inflation which is—in the strictly technical economic and not the publicly understood sense—non-transitory.

But instead of doing so, all that longer averaging windows do, implicitly, is drown pandemic transitory price rises in a large number of other observations. That is clearly inadequate to the task at hand.

Better by far is to correct that AIT conceptual mistake by calculating the inflation target adjustment on the trimmed-mean PCE, not the PCE. That procedure excludes outlier-cum-transitory price shocks from the target adjustment via the trimming process, and yields the following target table now:

Average Trimmed-Mean PCE Target, in percent

Averaging window centred on
November 2021 May 2022
H1 outturn H2 target H1 outturn H2 target
Four-year window 2.0 2.0 2.2 1.6
Six-year window 2.0 2.0 2.1 1.7
Ten-year window 1.9 2.1 2.0 1.9

May 2022 assumes that trimmed-mean PCE inflation is 3.5 percent until May 2022

Given that the trends in trimmed-mean PCE reflect the nucleus of consumer prices undistorted by outlier transitory shocks, these are coherent targets in contrast to those produced by AIT as is. And by eliminating outlier prices, a trimmed-mean PCE basis for target setting allows the time span of the window to be relatively short, a duration which is more congruent with its fundamental purposes.

These proposals do not require abandonment of the PCE in favour of the trimmed mean PCE as the concept used for the formal inflation target—now set at 2 percent. The two measures are identical over any reasonably long forecast horizon. All that changes is that the FOMC would use the trimmed-mean measure to set the target adjustments from 2 percent instead of the untrimmed measure. But given its proposed central status, the publication schedule at the Dallas Fed which produces the PCE trimmed mean measure should be brought forward to coincide with publication on the monthly CPI.

Thus, the twin aims that AIT was designed to secure would be rendered consistent with appropriate adjusted inflation targets now in mid-pandemic. Furthermore, as this matter is conceptual rather than conjectural, the fix is not just for the duration of the pandemic but is permanent. And with the target thus appropriately set to avoid the policy overkill implied now by AIT as is, this leaves the questions of if, when, and how to raise the central target up to 3 or 4 percent to be fought another day.

Even so, severe technical uncertainties would remain concerning what exactly the FOMC should do to hit the correctly adjusted target now in the context of the pandemic.

All this involves three technical terms which, because it has to communicate with the general public, are now absolute anathema to the Federal Reserve: “transitory”, “error”, and “trimmed-mean”. But all the policy alternatives are worse, with the option of carrying on and pretending that none of this is happening being the worst option of all.

Despite the total radio silence on this central matter in debate among economists and in the FOMC minutes, both top nominees for the Fed should be asked to swallow hard and commit to fix it, now.