What isn’t transitory is the tight labor market. Long-term demographics show that we’re going to be in a labor shortage for 20 years. Fed policy, economic models, etc. won’t solve this. The high-tech sector may be at an inflection point but it will be the high-tech sector (AI, automation, high-end manufacturing) must take up the slack in the labor market. The age-participation numbers are already set in motion. Unless we significantly increase immigration (politically infeasible with a gridlocked, coin-operated Congress), the burden/challenge falls to high tech.
Japanese politics, like Germany’s, are corporatist. Both have worse labor force demos than the US but lower wage gains.
There will be a decade or two where labor force is declining but population still increasing with a greater population of seniors with a high propensity to consume and dissave. Could mean more tech automation (robots are the same price everywhere- issue is land and electricity prices, freight rail and highway networks, taxes, building regs- US ain’t so bad on many metrics….though we don’t build robots or machine tools anymore), more old people and non-college folks lured back into workforce or simply more imports and offshoring (Japan starting to go that route).
During the whole section about "how long it takes" I was rolling my eyes and whining "come ON! What about the Doh and Foerster paper!"
... And then you mentioned it lol
In general I think everyone (academics, laypeople, pundits) underestimate how often systems lack any kind of "long run average". We crave a kind of certainty and predictability in the world that often just isn't there.
I think there is an obvious reason that monetary policy acts much more quickly now. Many years of low interest rates have made a bigger share of the population into leveraged speculators so they are now much more sensitive to rate increases. If there is anything in economics that is predictable, it is that raising rates from zero up to 5% over 12 months in a highly leveraged economy will produce a recession.
Seems like inflation has fallen back to 3-4 pct. Fed wants it back at 2-3 pct. My guess is we don’t see the Fed cutting until we get several months in a row of 2 pct handle core numbers (ex housing, food, energy) and unemployment is back above 4 pct. If we get a surprise recession we could see cuts without a return to 2 pct core inflation, I suppose, but Fed won’t like that.
It appears from what I see and how you describe various economists, that everyone only thinks and uses "linear elastic" modeling. But almost all the phenomena discussed has a "time component". Springs, elastic push-react and pull react is leaving out viscous time constants.
Look at all these types of curves from stress relaxation, creep, etc
There is another methodology to consider in macroeconomic modeling. We test polymers with....not impulse but....cyclic testing.
We're not interested here, in prediction. We are interested in characterization. We want to see the phase angle, ie time.delay between an input and the output.
I can't get you farther here with my agednans degrades skills. But I believe that there are a lot of various important relationships that have different time constants as well as linear response.
My economics professor in college won a Nobel prize for re-inventing linear control for interest rates. I was taking control theory at the time too and remember thinking “you can’t be serious, what about literally everything else we know about model predictive control?” But alas, the balance between art and science in macroeconomics is difficult.
I apologize for this embarrassingly newbie question, but how does home affordability factor into economic measurements? I ask because a common sentiment I experience talking to day-to-day people is that the economy is awful / inflation is awful because homes are less affordable now than ever.
That’s a pretty nuanced question. Are you talking about renters, homebuyers, home builders in the market now, or homeowners?
For renting, there are multiple data points pointing to filling costs of rent. Rent is tricky because leases are usually locked in for a year plus so it takes some time for them to be repriced. But they seem to be falling:
For homebuyers: housing prices have fallen, but there is still a historically low supply of housing and property values are still above the pre-pandemic trend. The raises and interests rate have also squeezed buyers, new making mortgage payments more expensive.
For homeowners that bought before the rate hikes, they’re sitting in 30 year mortgages that have historically low interest rates, so they are likely to be fine. I’ve seen some attempts to calculate the associated costs to homeowners, but I’m just not convinced it’s a relevant metric considering they have locked in/refinanced at a pretty low late. The one exception is that you see a reduce churn in hosing because people want to hold on to their low mortgage rates but that’s eh.
Home builders are facing high interests rate which makes building pretty expensive. But the construction sector is a mess and I’m not sure what to make of the data. It’s too much notice and I think we’ll know in. 6 months.
TLDR
-Rents decreasing
-New mortgages more expensive for homebuyers
-For homeowners prior March 2022, interests rates, drop and home value doesn’t really matter unless they’re looking to sell
Great article and analysis of all the perspectives. I think the various lags are being seen in goods but services (wages) will take much longer. The Fed’s sufficiently restrictive level will be very interesting given the core and stickier inflation areas.
Already the DXY is cratering and loosening financial conditions and that isn’t what the Fed wants as it makes inflation harder to fight. Example: US import prices have increased again.
Doesn't it follow from the popularity of predictions based on the yield curve, which is inherently driven by expectations, that expectations are important? I think, based on my own PhD research (also with input from Miles Kimball, btw), that expectation formation itself has a lag because of information lags and measurement noise. Lags in expectation formation are shorter than the 2-3 years you mention -- more like the six-month range. Also, lags are not the same as pure delays -- they are smooth, so the other guy's position that "...there’s no way that inflation could be falling because of the Fed’s rate hikes, because rate hikes take over a year to have an effect" sounds wrong to me.
Re the Fed's policy:
By my reckoning, the Fed was a year late in raising rates. Had they been on time, they would not have needed to be so aggressive. As they dilly-dallied, I worried that they would not be aggressive enough when they finally got around to it. I was happy my worries in that regard turned out to be unfounded. And for what it's worth, I think they are correct to still be raising rates slowly, but it won't be long until they need to level off and then lower rates (I predict fairly aggressively) to catch inflation at or near target. By delaying for so long, the Fed got itself into a situation in which aggressive moves and precise timing are needed to pull off a win. That said, it seems to me they are doing just that.
They've already over-corrected, since inflation is below 4 per cent, which is the sensible target. If there is a recession, it will be associated with a further decline in inflation, to the point where the zero low bound on the funds rate will be hit yet again, and more QE is needed. Supposedly, this will enhance the credibility of the Fed
If the Fed changes inflation targets in the middle of an inflationary episode it could lead to a loss of credibility, so I think any change to 4% will have to wait until long after they've gotten it back down to 2%.
It would have been much better to raise the target gradually in the years before the pandemic. But, as I've suggested, another round of QE wouldn't do much for credibility either.
Science seems to progress by continually taking in new observations and creating new, more complex theories to account for anomalous observations using newly forged tools. Newton invented calculus for his physics. Newtonian physics was replaced by relativity and quantum mechanics, which used previously unused mathematical tools. The relative simplicity of Darwin's evolution has been replaced by more complex models based on previously unused biochemistry and biophysics. Economics in this decade has been complicated by the effects of a pandemic and a shooting war between developed nation-states with relatively evenly matched mechanized weapon systems. I'm no economist, but this post suggests that economics is ready for a new toolbox..
Exactly. Economics is a creation of humans and if you don't understand humans then you have no hope. Rational economic man 😂!!! Have they never met any real people?
That said, there is a basic mathematical/logical component to economics, but if you can't get the definitions and the measurements right then you won't get any useful answers for this bit either.
Worth noting that the average white collar worker's belief in a coming recession is likely driven largely by the salience of layoffs from prominent companies. Of course, you're right to be reading data instead of extrapolating the vibes on LinkedIn, but to your point, inflation and recessions are both self-fulfilling prophecies, and the vibes on LinkedIn *are* likely affecting the consumption decisions of a group that consumes a lot.
I think it’s important to remember that much like Twitter, most people are not on LinkedIn and most that are do not use it on a daily basis so I think you may be over-extrapolating from your specific experience using the site. Moreover, most layoffs have occurred in tech, which is an extremely small sector comparatively speaking. What you have is a very vocal minority who is making thing seem much worse than they are when your average person who works for a contractor is consuming just as much as they usually do.
Totally! And Noah's data show that consumption is fine overall. My point is just that since these things are expectations based, sometimes very vocal minorities can swing them. But you're right that I might be over weighting my personal experience and that the tech minority is so small that even their outsized vocality doesn't actually make a significant dent in perceptions
So much of the discussion of how inflation is discussed in commentary is very different from what I learned doing my economics major. I was taught that interest rates fight inflation by vacuuming money out of the economy. Interest rates raise because the fed sells bonds like crazy through open market operations. This both lowers the price of bonds (raises interest rates) and hoovers up dollars. Fewer dollars in the economy raises the value of dollars relative to goods and services, which means dollar denominated prices go down (inflation goes down). Is that wrong? Does this just not happen, or is the effect too small to be worth mentioning? I think talking through this either to explain why is wrong or why it's right would be a great subject for a blog post. Or maybe someone can just explain it to me here...
Yes, there often seems to be a shortage of facts when it comes to these discussions. In theory the Fed sells bonds like crazy, but I'm not sure this actually happens in practice. Once the Fed announces it's new interest rate I think the market pretty much accepts it so they don't have to force the issue. This means the reduced money supply in fact comes from fewer loans being made by commercial banks, a much slower process. The Fed balance sheet is down about 5% in the last 6 months, most I think from QT (not rolling over bonds when they expire rather than outright selling)
Why wouldn't it happen in practice? There's no big lever at the Fed labeled "interest rates," open market operations are the mechanism by which the Fed adjusts interest rates. Do bondholders just ask 5% less for their assets because the Fed said so?
The central bank only has to buy if the market trades outside the declared band for interest rates. They have virtually unlimited firepower so the market mostly meekly accepts their diktat. Draghi only spoke a few words ("we will do whatever it takes") during the Euro crisis and the threat was enough for the markets to lower longer term interest rates. It's only if the policy rate doesn't make economic sense that the market will test the central banks resolve e.g. Japan's intervention in their longer term bond market. They are actually having to buy up loads of bonds to keep rates low
There's one big reason that inflation is falling, and will continue to fall - the M2 money supply is DECLINING. M2 previously hadn't seen a YoY fall in at least 60+ years.
It was the MoM falls in M2 that gave me the confidence many months ago to call June as the peak in CPI growth & for drastically lower inflation ahead.
Given the extent of the deceleration in M2, the reverse of what happened post the COVID stimulus is playing out now (e.g. cratering durables prices, falling spot market rents, plunging freight rates and lower oil & gasoline prices).
In my 2023 outlook I forecast CPI ex-shelter to not only fall below 2%, but also a significant likelihood of outright YoY DEFLATION by June (my full report is available for free on my Substack).
The Fed continuing to tighten in the face of falling M2 risks a SEVERE recession. There's absolutely NO need for additional tightening, which only risks MORE inflation, not less. Why? Because plunging the economy into a severe recession risks a repeat of the TRILLIONS in stimulus that caused the current bout of high inflation. This is what risks a 70s style yo-yoing of inflation, not pulling back on what is ALREADY the most aggressive tightening campaign in over 40 years.
You are taking Havranek & Rusnak WAY out of context. They use *both* hump-shaped and increasing responses to central bank policy in their estimate, and their results vary by country, with the US having a best estimate of 42 months - 3 and a half years.
By the source you have cited, the Fed has drastically over-reacted to the inflation rate, and an implication is we're likely to get a long a painful recession.
I'm sorry, but no. That U.S. "best estimate" averages the two types of impulse responses, but the two are mutually exclusive. Also, the measured lag is to the full effect, not to any effect at all. The picture helps.
What isn’t transitory is the tight labor market. Long-term demographics show that we’re going to be in a labor shortage for 20 years. Fed policy, economic models, etc. won’t solve this. The high-tech sector may be at an inflection point but it will be the high-tech sector (AI, automation, high-end manufacturing) must take up the slack in the labor market. The age-participation numbers are already set in motion. Unless we significantly increase immigration (politically infeasible with a gridlocked, coin-operated Congress), the burden/challenge falls to high tech.
Interesting. Why haven't demographics pushed up Japanese wages yet?
Japanese politics, like Germany’s, are corporatist. Both have worse labor force demos than the US but lower wage gains.
There will be a decade or two where labor force is declining but population still increasing with a greater population of seniors with a high propensity to consume and dissave. Could mean more tech automation (robots are the same price everywhere- issue is land and electricity prices, freight rail and highway networks, taxes, building regs- US ain’t so bad on many metrics….though we don’t build robots or machine tools anymore), more old people and non-college folks lured back into workforce or simply more imports and offshoring (Japan starting to go that route).
During the whole section about "how long it takes" I was rolling my eyes and whining "come ON! What about the Doh and Foerster paper!"
... And then you mentioned it lol
In general I think everyone (academics, laypeople, pundits) underestimate how often systems lack any kind of "long run average". We crave a kind of certainty and predictability in the world that often just isn't there.
I think there is an obvious reason that monetary policy acts much more quickly now. Many years of low interest rates have made a bigger share of the population into leveraged speculators so they are now much more sensitive to rate increases. If there is anything in economics that is predictable, it is that raising rates from zero up to 5% over 12 months in a highly leveraged economy will produce a recession.
Great overview and analysis, thanks
Seems like inflation has fallen back to 3-4 pct. Fed wants it back at 2-3 pct. My guess is we don’t see the Fed cutting until we get several months in a row of 2 pct handle core numbers (ex housing, food, energy) and unemployment is back above 4 pct. If we get a surprise recession we could see cuts without a return to 2 pct core inflation, I suppose, but Fed won’t like that.
Love this.
Here is the Big Idea for you, mentioned it before
It appears from what I see and how you describe various economists, that everyone only thinks and uses "linear elastic" modeling. But almost all the phenomena discussed has a "time component". Springs, elastic push-react and pull react is leaving out viscous time constants.
https://en.m.wikipedia.org/wiki/Viscoelasticity
Look at all these types of curves from stress relaxation, creep, etc
There is another methodology to consider in macroeconomic modeling. We test polymers with....not impulse but....cyclic testing.
We're not interested here, in prediction. We are interested in characterization. We want to see the phase angle, ie time.delay between an input and the output.
I can't get you farther here with my agednans degrades skills. But I believe that there are a lot of various important relationships that have different time constants as well as linear response.
My economics professor in college won a Nobel prize for re-inventing linear control for interest rates. I was taking control theory at the time too and remember thinking “you can’t be serious, what about literally everything else we know about model predictive control?” But alas, the balance between art and science in macroeconomics is difficult.
https://en.wikipedia.org/wiki/Taylor_rule
I apologize for this embarrassingly newbie question, but how does home affordability factor into economic measurements? I ask because a common sentiment I experience talking to day-to-day people is that the economy is awful / inflation is awful because homes are less affordable now than ever.
Yep, it does factor in. Basically they try to estimate how much it would take to rent your own home, and that's called "owner's equivalent rent".
That’s a pretty nuanced question. Are you talking about renters, homebuyers, home builders in the market now, or homeowners?
For renting, there are multiple data points pointing to filling costs of rent. Rent is tricky because leases are usually locked in for a year plus so it takes some time for them to be repriced. But they seem to be falling:
https://twitter.com/justinwolfers/status/1613535842747678720?s=46&t=M9KeLFDjClFLMyTTbOUNiA
For homebuyers: housing prices have fallen, but there is still a historically low supply of housing and property values are still above the pre-pandemic trend. The raises and interests rate have also squeezed buyers, new making mortgage payments more expensive.
For homeowners that bought before the rate hikes, they’re sitting in 30 year mortgages that have historically low interest rates, so they are likely to be fine. I’ve seen some attempts to calculate the associated costs to homeowners, but I’m just not convinced it’s a relevant metric considering they have locked in/refinanced at a pretty low late. The one exception is that you see a reduce churn in hosing because people want to hold on to their low mortgage rates but that’s eh.
Home builders are facing high interests rate which makes building pretty expensive. But the construction sector is a mess and I’m not sure what to make of the data. It’s too much notice and I think we’ll know in. 6 months.
TLDR
-Rents decreasing
-New mortgages more expensive for homebuyers
-For homeowners prior March 2022, interests rates, drop and home value doesn’t really matter unless they’re looking to sell
-builders: 🤷🏾♂️
Thank you so much for that detailed response.
Great article and analysis of all the perspectives. I think the various lags are being seen in goods but services (wages) will take much longer. The Fed’s sufficiently restrictive level will be very interesting given the core and stickier inflation areas.
Already the DXY is cratering and loosening financial conditions and that isn’t what the Fed wants as it makes inflation harder to fight. Example: US import prices have increased again.
I enjoyed the Wile E. Coyote reference so looked it up. Krugman used it in 2007 in the context of dollar mis-pricing: https://archive.nytimes.com/krugman.blogs.nytimes.com/2007/09/20/is-this-the-wile-e-coyote-moment/
Re the importance of expectations:
Doesn't it follow from the popularity of predictions based on the yield curve, which is inherently driven by expectations, that expectations are important? I think, based on my own PhD research (also with input from Miles Kimball, btw), that expectation formation itself has a lag because of information lags and measurement noise. Lags in expectation formation are shorter than the 2-3 years you mention -- more like the six-month range. Also, lags are not the same as pure delays -- they are smooth, so the other guy's position that "...there’s no way that inflation could be falling because of the Fed’s rate hikes, because rate hikes take over a year to have an effect" sounds wrong to me.
Re the Fed's policy:
By my reckoning, the Fed was a year late in raising rates. Had they been on time, they would not have needed to be so aggressive. As they dilly-dallied, I worried that they would not be aggressive enough when they finally got around to it. I was happy my worries in that regard turned out to be unfounded. And for what it's worth, I think they are correct to still be raising rates slowly, but it won't be long until they need to level off and then lower rates (I predict fairly aggressively) to catch inflation at or near target. By delaying for so long, the Fed got itself into a situation in which aggressive moves and precise timing are needed to pull off a win. That said, it seems to me they are doing just that.
Yeah, 6 months for expectation formation sounds about right to me!
This is timely and extremely elucidating. Thanks, Noah. I'm going to make sure Drum sees this (though he probably has already).
Typo in the paragraph after the “median cpi” graph. “where you don’t nearly as much worry about inflation anymore.”
Thanks, fixed!
They've already over-corrected, since inflation is below 4 per cent, which is the sensible target. If there is a recession, it will be associated with a further decline in inflation, to the point where the zero low bound on the funds rate will be hit yet again, and more QE is needed. Supposedly, this will enhance the credibility of the Fed
If the Fed changes inflation targets in the middle of an inflationary episode it could lead to a loss of credibility, so I think any change to 4% will have to wait until long after they've gotten it back down to 2%.
It would have been much better to raise the target gradually in the years before the pandemic. But, as I've suggested, another round of QE wouldn't do much for credibility either.
Science seems to progress by continually taking in new observations and creating new, more complex theories to account for anomalous observations using newly forged tools. Newton invented calculus for his physics. Newtonian physics was replaced by relativity and quantum mechanics, which used previously unused mathematical tools. The relative simplicity of Darwin's evolution has been replaced by more complex models based on previously unused biochemistry and biophysics. Economics in this decade has been complicated by the effects of a pandemic and a shooting war between developed nation-states with relatively evenly matched mechanized weapon systems. I'm no economist, but this post suggests that economics is ready for a new toolbox..
Exactly. Economics is a creation of humans and if you don't understand humans then you have no hope. Rational economic man 😂!!! Have they never met any real people?
That said, there is a basic mathematical/logical component to economics, but if you can't get the definitions and the measurements right then you won't get any useful answers for this bit either.
Worth noting that the average white collar worker's belief in a coming recession is likely driven largely by the salience of layoffs from prominent companies. Of course, you're right to be reading data instead of extrapolating the vibes on LinkedIn, but to your point, inflation and recessions are both self-fulfilling prophecies, and the vibes on LinkedIn *are* likely affecting the consumption decisions of a group that consumes a lot.
I think it’s important to remember that much like Twitter, most people are not on LinkedIn and most that are do not use it on a daily basis so I think you may be over-extrapolating from your specific experience using the site. Moreover, most layoffs have occurred in tech, which is an extremely small sector comparatively speaking. What you have is a very vocal minority who is making thing seem much worse than they are when your average person who works for a contractor is consuming just as much as they usually do.
Totally! And Noah's data show that consumption is fine overall. My point is just that since these things are expectations based, sometimes very vocal minorities can swing them. But you're right that I might be over weighting my personal experience and that the tech minority is so small that even their outsized vocality doesn't actually make a significant dent in perceptions
So much of the discussion of how inflation is discussed in commentary is very different from what I learned doing my economics major. I was taught that interest rates fight inflation by vacuuming money out of the economy. Interest rates raise because the fed sells bonds like crazy through open market operations. This both lowers the price of bonds (raises interest rates) and hoovers up dollars. Fewer dollars in the economy raises the value of dollars relative to goods and services, which means dollar denominated prices go down (inflation goes down). Is that wrong? Does this just not happen, or is the effect too small to be worth mentioning? I think talking through this either to explain why is wrong or why it's right would be a great subject for a blog post. Or maybe someone can just explain it to me here...
Yes, there often seems to be a shortage of facts when it comes to these discussions. In theory the Fed sells bonds like crazy, but I'm not sure this actually happens in practice. Once the Fed announces it's new interest rate I think the market pretty much accepts it so they don't have to force the issue. This means the reduced money supply in fact comes from fewer loans being made by commercial banks, a much slower process. The Fed balance sheet is down about 5% in the last 6 months, most I think from QT (not rolling over bonds when they expire rather than outright selling)
Why wouldn't it happen in practice? There's no big lever at the Fed labeled "interest rates," open market operations are the mechanism by which the Fed adjusts interest rates. Do bondholders just ask 5% less for their assets because the Fed said so?
The central bank only has to buy if the market trades outside the declared band for interest rates. They have virtually unlimited firepower so the market mostly meekly accepts their diktat. Draghi only spoke a few words ("we will do whatever it takes") during the Euro crisis and the threat was enough for the markets to lower longer term interest rates. It's only if the policy rate doesn't make economic sense that the market will test the central banks resolve e.g. Japan's intervention in their longer term bond market. They are actually having to buy up loads of bonds to keep rates low
There's one big reason that inflation is falling, and will continue to fall - the M2 money supply is DECLINING. M2 previously hadn't seen a YoY fall in at least 60+ years.
It was the MoM falls in M2 that gave me the confidence many months ago to call June as the peak in CPI growth & for drastically lower inflation ahead.
Given the extent of the deceleration in M2, the reverse of what happened post the COVID stimulus is playing out now (e.g. cratering durables prices, falling spot market rents, plunging freight rates and lower oil & gasoline prices).
In my 2023 outlook I forecast CPI ex-shelter to not only fall below 2%, but also a significant likelihood of outright YoY DEFLATION by June (my full report is available for free on my Substack).
The Fed continuing to tighten in the face of falling M2 risks a SEVERE recession. There's absolutely NO need for additional tightening, which only risks MORE inflation, not less. Why? Because plunging the economy into a severe recession risks a repeat of the TRILLIONS in stimulus that caused the current bout of high inflation. This is what risks a 70s style yo-yoing of inflation, not pulling back on what is ALREADY the most aggressive tightening campaign in over 40 years.
You are taking Havranek & Rusnak WAY out of context. They use *both* hump-shaped and increasing responses to central bank policy in their estimate, and their results vary by country, with the US having a best estimate of 42 months - 3 and a half years.
By the source you have cited, the Fed has drastically over-reacted to the inflation rate, and an implication is we're likely to get a long a painful recession.
I'm sorry, but no. That U.S. "best estimate" averages the two types of impulse responses, but the two are mutually exclusive. Also, the measured lag is to the full effect, not to any effect at all. The picture helps.