Inflation was obviously transitory
There seems to be a developing consensus that transitory inflation was a myth that was proven wrong because (1) inflation was high for longer than anticipated, (2) it led to a permanent rise in the price level, and (3) it took rate hikes to bring it down.
I have written about this quite a bit. My response to points 2 and 3 is that inflation came down without rate hikes, and, to the contrary, the permanent rise in the price level was due to the fact that Fed policy kept prices from deflating. The Fed was loose as inflation subsided, not tight. And, since transitory inflation lasted longer than anticipated, that was the right thing to do.
One question at the center of the issue is: Should the Fed aim to counteract inflation that is known to be temporary? To me, the answer is clearly, “No”. And the reason the answer is clear is that there is little gray area on this question in real time. If inflation is monetary in nature, and requires a response from a central bank that targets interest rates, it will attain an unstable equilibrium. Inflation will pull the neutral target interest rate higher until it is reversed. If it doesn’t require a response, it won’t.
This is reflected in a practical sense in the commonly heard concern about things like wage-price spirals. And for a rate-targeting central bank, it is true in an arithmetic sense, since rising inflation mathematically raises the neutral nominal interest rate. This is a frequent problem for the Fed when policy changes are tardy and they end up chasing a moving target up or down. When the Fed had an inflationary bias in the 1970s, it tended to chase rates up when inflation accelerated. During the Great Moderation, it tended to chase rates down leading into recessions. In late 2007 and early 2008, for instance, they cut the target rate from 5.25% to 2%, but they weren’t injecting cash into the market. They weren’t increasing their holdings of Treasuries.
I have written about how the norm of using trailing 12 month inflation has created confusion because when inflation abruptly ended, it took a year for it to age out of the measure. This made it look like the Fed was battling to get inflation down in the spring of 2023 when inflation had been back to normal since July 2022.
I’m going to walk through the timeline of inflation and policy rates again, to review anti-transitory claims. And, here, I’m going to use rates, expected inflation (estimated by the Cleveland Fed), and (instead of trailing inflation) forward inflation. Forward inflation, after all, is the inflation that matters for interest rates.
The Fed’s target is 2% PCE inflation, which equates to a bit more than 2% CPI inflation. In March 2022, 3 month Treasuries (which mostly reflect the Fed target rate and short-term expectations of its target rate) were paying 0.44%, expected inflation was 3.1%, and forward 3 month inflation turned out to be 10%.
By June, 3 month Treasuries were paying 1.5%, expected inflation was up to 4.2%, and forward 3 month inflation turned out to be 3.0%
So, we can infer that:
The market had expected inflation to decline sharply since early 2021, and it continued to expect it.
Inflation did decline sharply.
Even as inflation declined sharply, the market was losing faith in that prospect. Expected 1-year inflation was higher in July 2022 after inflation normalized than it had been when the Fed first raised its target rate in March 2022. Even today expected inflation is higher than it was when the Fed first raised its target rate.
In March 2022 - the last month before forward 3-month inflation finally started dropping quickly back to normal - the realized real interest rate on 3-month Treasuries was negative 9.5%.
I don’t see how the idea that the Fed tightened us out of this inflation can remotely withstand these basic facts. Real interest rates were deeply negative and inflation expectations were rising when inflation normalized.
Of course the actual target interest rate had little to do with the end of transitory inflation. Nobody has ever claimed that a negative 9.5% target rate is contractionary. But, interest rate targeting monetary policy has both long and variable lags and forward expectations. There is nothing it can’t explain. And, so, here, some attribute the abrupt end of inflation to the Fed’s commitment to rate hikes in the future.
But, forward rates topped out at about 3% when inflation normalized. There is no direct evidence of expected rates that would have been nearly high enough to reverse 10% inflation.
So, one might try to salvage the anti-transitory story by saying, “Well, why can’t it be both? Of course there was some temporary inflation from federal stimulus, Covid waves, and the Ukraine War. There was going to be some decline in inflation, but it was Fed tightening that brought it all the way back.”
OK. Let’s ignore the facts above and consider this story. What is a reasonable hawkish rate stance? Maybe something like anything more than 1% above inflation? Like, if inflation is running at 3%, a 4% rate target might be high enough to pull the trend down?
Well, the target inflation rate is 2%. In March 2022, the 3-month interest rate was 0.4% and the 1-year interest rate was 1.3%. One-year expected inflation was 3.1%. In July 2022, the 3-month interest rate was 2.2% and the 1-year rate was 3.2%. So, even taking expectations into account, for the target rate to be considered contractionary, expected inflation would have to be negligible in March or less than 2% by June and July when inflation broke. Expected 1-year inflation was 3% to 4%. Measured against either actual inflation or the much lower expected inflation rate, the Fed target rate and future expected target rates were negative.
Inflation dropped by about 7% after June 2022. How should we account for that? By my math, about 130% of the decline was transitory and Fed policy mitigated the additional 30%. Fed policy was loose at the time.
I suppose you can construct some sort of double causality. Inflation declined because the power of expectations is so strong, forward rates didn’t even have to rise to the level of inflation in order for markets to know the Fed was going to bring inflation down. Like never having to use a bomb if your bomb is big enough.
I do think expectations are key - but not monetary policy expectations. There was a strong market presumption that the inflation was transitory! Because it was! The effectiveness of a rate-targeting monetary policy depends on this.
Most lending happens with long durations. I think the average duration of marketable debt is around 6 years. So, since the inflation was transitory, it didn’t matter that much what the short term rate was, and it wouldn’t make sense to force the short term rate up to match the temporary inflation, because the interest rates and inflation expectations six years from now are what are driving borrowing and spending decisions. The short term target rate matters a lot when inflation is permanent because it is the result of the monetary stance and the short end of the curve is a coherent part of a standard yield curve.
That’s also why the too-tight policy the Fed has today isn’t as bad as it could be. Because the market expects it to reverse in the next year, so spending and borrowing decisions are relatively unaffected by it. Nobody is financing a semiconductor plant on overnight repos.
The reason “expectations” seem like the important element in the end of inflation is because they were! Inflation was expected to be transitory. And it was! It has nothing to do with expected Fed policy. Fed policy was loose when inflation reversed, which kept the end of transitory inflation from diving into disruptive deflation.
This is one of countless ways that I think the “Upside-down Capital Asset Pricing Model” can be helpful. The traditional way of viewing monetary policy as a manipulation of interest rates to stimulate borrowing mostly creates confusion. When the real target interest rate was negative 9%, it was a little loose, but it wasn’t loose enough to make the transitory inflation become unstable. Money supply was pretty flat in early 2022. There were probably some capital markets that got a bit of a boost from such low interest rates. But that’s probably about where rates would have been if we didn’t communicate monetary policy with a rate target. It’s not like the Fed was buying trillions in new bonds.
There are a few economists who understand that the Fed was loose. Marcus Nunes and Scott Sumner get it. (Please comment if you know of others who understand that the Fed was loose in late 2022, who should be mentioned here.) Marcus, Scott, and I wouldn’t agree on everything. Scott, for instance, thinks they let things get too hot. I think that one can reasonably support the policy trajectory in which the Fed guided us back to a 5% NGDP growth path. But, at least they have a coherent point of view about Fed policy that can be discussed.
How do you have a quantifiable debate about Fed policy that can be called tight with a negative 9% real target rate? Has anyone done much more than just hand-waving at the expectations channel? It reminds me of 2008, when it had been decided that a bubble had led to a (yet to be realized, but supposedly inevitable) real estate correction. Would it be a 5% correction? 10%? 20%? 30%? It ended up at 30%, so the canon that became the intellectual water economists now swim in was that this proved that a 30% correction had been inevitable. Any result in home prices after 2007 would have confirmed the “model”. The entire model was a residual. The conclusion had been presumed and any scale would confirm it.
Fed policy brought down inflation in July 2022. It could have come down to 6%, 4%, 2%, or less. What caused it to decline to 2%? A residual term in our mental equation called “expectations” caused it to decline to 2%.
It’s so weird to me how macroeconomists are constantly bickering about rate targets. Is a 1% hike enough? Will it take 1.25%? Debating a dozen tweaks to the Taylor Rule adjusting rate changes by a few basis points. And, suddenly, in mid-2022, there’s an immediate 7% inflation drop after a couple small hikes and everyone’s just like, “Well, sure. It was rate hike expectations.”
If the sensitivity of inflation to rate changes can be this pliable, how can we know anything? What if they had moved earlier and rates had gotten to 4% by June? Would it have plunged us to deep deflation? Or would the hikes above 2% have suddenly become just as impotent as the first 1% was powerful?
What if they had kept the target rate at 0%? If a negative 9% target rate suddenly brought inflation down to 3%, maybe it would have just come down more slowly if the rates were never hiked at all. I mean, if a negative 9% rate can be contractionary, why can’t negative 11%?
If Fed hikes were the key to falling inflation, it’s like an “everything, everywhere, all the time” monetary policy. We could lower the target 0.5% and suddenly have 20% inflation. How could we know it wouldn’t happen? Have the anti-transitory folks lost all faith in our ability to model interest rate targets? Or, once the transitory discussion dies down, will they go back to debating how many basis points will be required to fine tune next quarter’s employment growth and residential investment?
The thing is, I don’t think a lot of economists were calling for a 12% target rate. I think in a lot of cases, they called for 3% to 5%. Which, as I argued above, basically presumes that most of the inflation would have to be transitory for those rates to be effective.
Finally, in Figure 2 I use my preferred “CPI less Shelter” measure. As I frequently mention, the CPI shelter (mostly tracking housing rents) component has known lags which have made it inaccurate in these conditions. One way to adjust for that is to replace it using estimates from Zillow or some other rent-tracking service. But, really, the short-term trends of market rents track quite closely with non-rent inflation, so just taking shelter out of the CPI gives a similar result for the second derivate (changes in the rate of inflation). It tends to run about 1% lower than the aggregate CPI that includes shelter (the first derivative of prices - the rate of inflation). For reasons readers have probably heard from me before, I think that’s a better measure of monetary policy anyway. Rising land rents shouldn’t be confused with monetary inflation, especially when the vast majority of them are hedged through homeownership. Housing is a much smaller portion of PCE than of CPI. That’s a big reason why it tends to run lower than CPI. So, in a way, the Fed does target an inflation rate closer to my “CPI less shelter” preference.
Anyway, the story here is so clear (to me anyway) that it doesn’t really require using the right inflation measure. But, if you use it, the story is even more extreme. Forward 3-month inflation was really 14% in March 2022 and by June it was negative 2%. It remained negative through the end of 2022 before moving back up around the 2% target range throughout 2023.
So, the target policy interest rate in March 2022 was negative 13% and inflation quickly dropped by 16%. Even with the relatively loose stance, the Fed didn’t completely counter the deflationary reversal of transitory inflation in late 2022.
It’s a good thing that, as transitory inflation became extended, they maintained their somewhat loose posture. At some level of deflation, we might have tempted recessionary risks.
PS. In case you’re confused about the charts, I included a one-year forward inflation measure in the figures that I never mentioned in the text. I have left it in, in case anyone finds it useful. It is intended to be paired with the 1-year forward expected inflation. Sort of an expected vs. realized comparison. But, I find it difficult to construction a discussion of it as a reflection of policy over time. So I didn’t discuss it. But, you can see, in any given month, in the figures, how expected inflation and realized inflation compared.