Warsh spent part of his testimony hunting for a superior inflation gauge — trimmed means, medians, a new measure of what a big-box retailer charges for a good.
Admirable concatenation of metaphors contrasting micro and macro. But how do we know that it does not require NGDP growing at 6% to iron out the wrinkles in relative prices arising from Iran, from the tariffs from the deportations from the extreme sector specificity of investment in AI rollout? Even if we knew that 4% is optimal inflation-minimizing/real income-maximizing rate over the long term with normal levels of shocks, that tells us nothing about what the optimal inflation-minimizing/real income-maximizing rate in 2026 Q2 is.
Who is to say that hints may not be found in comparing headline with indexes that capture price (none) movement in sticky price sectors.
And recall that fever, the body's "temporarily over-target" temperature, is adaptive; it's cells can tolerate the extreme heat more than many bacterial.
Fair challenge, and the fever line is a good jab — granted, a temporary overshoot can be adaptive. That's precisely the case for a level target: it permits the fever and mandates the recovery, unlike a rate target that ratifies whatever temperature it wakes up with.
On the deeper question — how do we know 6% NGDP growth isn't what's needed to grease this quarter's relative-price adjustments? — the level-targeting answer is that we don't need to know, and shouldn't guess. Relative prices (Iran's oil, tariffs, AI-sector wages) can adjust within a stable spending path; what they need is a predictable nominal environment, not extra nominal fuel. Pouring in more NGDP to "iron out wrinkles" is exactly the accommodation that turned relative shocks into general inflation in the 1970s and 2021. And the epistemic point cuts the other way: if no one can compute the optimal rate for 2026 Q2 — and no one can — the argument for a fixed, pre-announced path gets stronger, not weaker. Discretion needs the knowledge you've just shown we lack.
Where we agree: sticky-price indexes are useful diagnostics for how the shock is splitting. My objection is to promoting the thermometer into the target.
One difference is basically rule v discression. If we have a rule, NGDPLT is better than AIT But is it better than PLT? And is it better than FPLT = (?) FNGDPLT And how do you derive the LT from first principles?
When there is a big upwards cost shock, it is an overwhelming bet that the corresponding price will go up a lot, especially if the industry is running near flat-out whether or not it is characterized by market power.
When there is a big negative cost shock, more often than not the price will not go down by much. Instead, rather than cut their prices and continue their provision, industries that are not with market power will cut back on provision in order to maintain close to per-unit revenue.
Thus in a world in which there are a lot of cost shocks, whether transitory or more permanent due to resource depletion or technological progress, while have an upward gap between the average price increase and the median. Or so it will be as long as either the transitory or resource-depletion permanent cost shock is confined to a small part of the economy or the technology shock is of the leading sector-focused Schumpeterian creative-destruction type.
Some thought would then lead one to the conclusion that there is then something like a natural rate of average inflation: the median inflation rate consistent with "effective price stability", whatever that may be, plus whatever wedge balances the equities between the informational costs of nominal price instability on the one hand and the smooth and proper greasing of structural change via the appropriate relative-price signals. If one wanted a market signal-based monetary policy that took account of these considerations, one could do worse than look at the median rate of price increase.
Not, however, to say that Kevin Warsh thinks along these lines. Instead, he seems to have some view that median inflation gets you a better forecast of future inflation than either headline or core inflation does. That has not been demonstrated to me.
> **Marcus Nunes**: A Better Thermometer for the Wrong Fever <https://marcusnunes.substack.com/p/a-better-thermometer-for-the-wrong>: The chairman of the Federal Reserve, one meeting into the job, having just staked the institution’s credibility on defeating inflation, has told the world that his problem is finding the right price index. Not the right policy. The right thermometer. This is the oldest error in monetary economics wearing a data-science costume, and it deserves to be named, because a great deal of bad policy has been built on it.... The trimmed mean is a clumsy, backward attempt to recover a nominal-demand signal from price data.... And there is a series that already is that signal, measured directly, no trimming required: nominal GDP..... Here is the part that turns a conceptual error into a policy danger. A Fed that reads inflation off a better price index will, by construction, respond to the wrong things...
======
**DELONG’S GRASPING REALITY: Trying to make my readers—and myself—smarter. I think I am a go-to source to understand things economic in the past and in the present. Too online since 1995. Subscribe! <https://graspingreality.substack.com/subscribe>Currently featuring**:
"Some thought would then lead one to the conclusion that there is then something like a natural rate of average inflation: the median inflation rate consistent with "effective price stability." [N.B. getting to this required cosideration of adjustment of relative prices among diverse goods, services, and factors of roduction, not just real wage of one undifferentiated input into one undefferentiated output .]
DeLong produces the theoretical rational for (forward-looking) Average Inflation Targeting (AIT) at a non-zero rate. Allow for periods of extraordinary large shocks and and you have the rationale for Flexible Averate Inflation Targeting (FAIT). Add an estimate of maximum real income growth and AIT => NGDPLT and FAIT => FNGDPLT.
Admirable concatenation of metaphors contrasting micro and macro. But how do we know that it does not require NGDP growing at 6% to iron out the wrinkles in relative prices arising from Iran, from the tariffs from the deportations from the extreme sector specificity of investment in AI rollout? Even if we knew that 4% is optimal inflation-minimizing/real income-maximizing rate over the long term with normal levels of shocks, that tells us nothing about what the optimal inflation-minimizing/real income-maximizing rate in 2026 Q2 is.
Who is to say that hints may not be found in comparing headline with indexes that capture price (none) movement in sticky price sectors.
And recall that fever, the body's "temporarily over-target" temperature, is adaptive; it's cells can tolerate the extreme heat more than many bacterial.
Fair challenge, and the fever line is a good jab — granted, a temporary overshoot can be adaptive. That's precisely the case for a level target: it permits the fever and mandates the recovery, unlike a rate target that ratifies whatever temperature it wakes up with.
On the deeper question — how do we know 6% NGDP growth isn't what's needed to grease this quarter's relative-price adjustments? — the level-targeting answer is that we don't need to know, and shouldn't guess. Relative prices (Iran's oil, tariffs, AI-sector wages) can adjust within a stable spending path; what they need is a predictable nominal environment, not extra nominal fuel. Pouring in more NGDP to "iron out wrinkles" is exactly the accommodation that turned relative shocks into general inflation in the 1970s and 2021. And the epistemic point cuts the other way: if no one can compute the optimal rate for 2026 Q2 — and no one can — the argument for a fixed, pre-announced path gets stronger, not weaker. Discretion needs the knowledge you've just shown we lack.
Where we agree: sticky-price indexes are useful diagnostics for how the shock is splitting. My objection is to promoting the thermometer into the target.
One difference is basically rule v discression. If we have a rule, NGDPLT is better than AIT But is it better than PLT? And is it better than FPLT = (?) FNGDPLT And how do you derive the LT from first principles?
When there is a big upwards cost shock, it is an overwhelming bet that the corresponding price will go up a lot, especially if the industry is running near flat-out whether or not it is characterized by market power.
When there is a big negative cost shock, more often than not the price will not go down by much. Instead, rather than cut their prices and continue their provision, industries that are not with market power will cut back on provision in order to maintain close to per-unit revenue.
Thus in a world in which there are a lot of cost shocks, whether transitory or more permanent due to resource depletion or technological progress, while have an upward gap between the average price increase and the median. Or so it will be as long as either the transitory or resource-depletion permanent cost shock is confined to a small part of the economy or the technology shock is of the leading sector-focused Schumpeterian creative-destruction type.
Some thought would then lead one to the conclusion that there is then something like a natural rate of average inflation: the median inflation rate consistent with "effective price stability", whatever that may be, plus whatever wedge balances the equities between the informational costs of nominal price instability on the one hand and the smooth and proper greasing of structural change via the appropriate relative-price signals. If one wanted a market signal-based monetary policy that took account of these considerations, one could do worse than look at the median rate of price increase.
Not, however, to say that Kevin Warsh thinks along these lines. Instead, he seems to have some view that median inflation gets you a better forecast of future inflation than either headline or core inflation does. That has not been demonstrated to me.
> **Marcus Nunes**: A Better Thermometer for the Wrong Fever <https://marcusnunes.substack.com/p/a-better-thermometer-for-the-wrong>: The chairman of the Federal Reserve, one meeting into the job, having just staked the institution’s credibility on defeating inflation, has told the world that his problem is finding the right price index. Not the right policy. The right thermometer. This is the oldest error in monetary economics wearing a data-science costume, and it deserves to be named, because a great deal of bad policy has been built on it.... The trimmed mean is a clumsy, backward attempt to recover a nominal-demand signal from price data.... And there is a series that already is that signal, measured directly, no trimming required: nominal GDP..... Here is the part that turns a conceptual error into a policy danger. A Fed that reads inflation off a better price index will, by construction, respond to the wrong things...
======
**DELONG’S GRASPING REALITY: Trying to make my readers—and myself—smarter. I think I am a go-to source to understand things economic in the past and in the present. Too online since 1995. Subscribe! <https://graspingreality.substack.com/subscribe>Currently featuring**:
* The Fourteen-Lion Parade <https://braddelong.substack.com/p/fourteen-lion-parade>
* DRAFT: Notes on Human Division-of-Labor Globalization from Stone Tools to the Assembly Line: -750000 to 1945 <https://braddelong.substack.com/p/draft-notes-on-human-division-of>
* Consequences of the Revolutions of 1848: Élite Recognition that "If Everything Is Going to Stay the Same, Everything Has to Change..." <https://braddelong.substack.com/p/draft-consequences-of-the-revolutions>
======
<https://marcusnunes.substack.com/p/a-better-thermometer-for-the-wrong>
"Some thought would then lead one to the conclusion that there is then something like a natural rate of average inflation: the median inflation rate consistent with "effective price stability." [N.B. getting to this required cosideration of adjustment of relative prices among diverse goods, services, and factors of roduction, not just real wage of one undifferentiated input into one undefferentiated output .]
DeLong produces the theoretical rational for (forward-looking) Average Inflation Targeting (AIT) at a non-zero rate. Allow for periods of extraordinary large shocks and and you have the rationale for Flexible Averate Inflation Targeting (FAIT). Add an estimate of maximum real income growth and AIT => NGDPLT and FAIT => FNGDPLT.
:-)