A note on levels and growth rates

Some commenters have pushed back on my “no tight money” claims by pointing to a slowdown in NGDP growth over the past year. I am not persuaded.

Under any sort of rational policy regime, you should try to come back to the trend line after a steep decline caused by special factors. NGDP fell extremely sharply in mid-2020 (due to very special factors), and hence it was appropriate that NGDP rose extremely rapidly over the following year. That’s the implication of level targeting, but also average inflation targeting and even the Fed’s vague “dual mandate”.

But once NGDP got back near the old trend line, it was appropriate to slow NGDP growth to a rate consistent with the Fed’s 2% target, which implies an NGDP growth rate of roughly 3.5%. Instead, NGDP growth has averaged 9% over the past year. Even if that’s a bit slower than the rate during the bounce back from Covid, policy is actually becoming more expansionary in the sense that we are now moving further and further above the optimal target path, instead of merely recovering to the previous trend line (which was appropriate). It’s not enough to think in terms of growth rates, you also need to consider levels.

I understand that concepts such as easy and tight money are subjective, so let me make my point using different terminology. I believe that the Fed’s policy error over the past 12 months has been even worse than the mistakes made in 2021. Policy in 2021 was too expansionary (at least late in the year), but the policy in 2022 is even worse, far too expansionary.

It’s possible that policy has recently become a bit less expansionary, but it’s certainly not a tight money policy by any reasonable definition. Forecasts for 4th quarter NGDP growth remain quite high. Every month of overshooting makes the subsequent correction even more painful. Stopping high inflation in 2023 will be more painful (or less effective) than it would have been if we’d bitten the bullet and brought NGDP growth down more sharply this year. Nominal wage growth has developed a lot of upward momentum, making it much harder to control.



49 Responses to “A note on levels and growth rates”

  1. Gravatar of George George
    18. November 2022 at 14:21

    Wouldn’t the last two years be itself a new ‘trend’ (of inflation tax without representation)? How long of a period must transpire before the definition of ‘trend’ is updated?

    If a ‘trend’ of 3.5% is caused by Central Bank whim, why is a Central Bank whim of 9% defined as something other than ‘trend’?

    It’s clear that it’s subjective turtles all the way down. The reason 3.5% is called a ‘trend’ is the same reason 9% is also a ‘trend’.

    Why EXACTLY is 9% ‘too high’ while 3.5% is ‘just right’ while 1% is ‘too low’, if each number ARE THEMSELVES SUBJECTIVELY DETERMINED, and can each be repeated year after year by Central Bank whim?

    The problem, it is clear, isn’t at the ‘level’ of any whim, excuse me, of any ‘target’. The problem is what logic is driving changes. If 9% was the chosen whim year after year after year, then because of the absence of any ECONOMY determined logic for why 9% is taking place, the whole ‘level NGDP targeting’ logic MUST output that 9% year after year is ‘just right’, because it is itself a ‘trend’.

    For there would presumably not be any ‘long and variable lags’, there would presumably not be any interest rate manipulation caused distortions in the capital structure, there would presumably be no ‘distortions’ as the ‘efficient market’ would ‘ingest’ a 9% trend.

    Notice how the definitions of ‘too low’ and ‘too tight’ being presented by the site owner, are circular definitions that themselves depend on subjectively determined definitions of ‘too tight’ and ‘too low’?

    I am not advocating for any of this, just asking to test the logic, why does the Central Bank NEED to ‘revert’ BACK to less than 9% in the first place? Why not 9% going forward? If that is assumed as the new ‘trend’, then that in turn because of the circular logic would compel us to define the current 9% trend as ‘just right’.

    The only argument against a permanent 9% vs a permanent 3.5%, is subjectivity. Site owner has to admit this because the entire ‘NGDP level targeting’ logic is totally absent of any consideration of distortions caused by ANY ‘NGDP level targeting’ of inflation.

    In other words, this doesn’t have to be deliberate or intentional but it’s true: the logic underlying tyrannical dictatorships, namely the subjective whims of those with government level powers who overrule all other people’s subjectivity, is the same exact logic underlying ‘NGDP level targeting’. The subjectivity of those in power positions of government overrule all other people’s subjectivity.

    There is no connection, no link, no logic interface with, free enterprise, where INTER-subjective results of economic cooperation establish common ground, i.e. objective, economic realities that convey real world economic data of the demand AND SUPPLY for the most marketable commodity, i.e. money.

    Again it’s subjectivity turtles ALL the way down. Every attempt to escape a subjective definition fails. Site owner proves this in this post. That’s why the same subjectivity can only be ‘restated’ with the same subjective logic. To wit:

    “I understand that concepts such as easy and tight money are subjective, so let me make my point using different terminology. I believe that the Fed’s policy error over the past 12 months has been even worse than the mistakes made in 2021. Policy in 2021 was too expansionary (at least late in the year), but the policy in 2022 is even worse, far too expansionary.”

    Translation: I understand that concepts such as easy and tight money are subjective, so here’s more subjective conceptualizations to really drive that point home.

    See the inconsistency yet? While 9% is defined as ‘too high’…it is ALSO defined as ‘just right’ and that is said when the subjectivity leads to more subjectivity and 9% is assumed or ‘conceptualized’ as a new ‘trend’, because muh level target is what matters. But then a 9% ‘trend’ as ‘just right’ would make 3.5% suddenly ‘too tight’, because muh level target is what matters.

    THen jUsT PiCk aNy LeVeL tARgEt aNd sTiCk WiTh iT!

    How can anyone do that when they’re simultaneously being told that any number is both too high and too low and just right, depending on one’s SUBJECTIVE assessment of current and expected future ‘levels’ and subjective definition of ‘trend’? The subjective NGDP level logic permits anyone to conclude that 50 years of 3.5% is ‘too tight’, if the ‘just right’ definition is 9%.

    What’s really funny to watch is the dialectic logic faith being used by the site owner, where there is a simultaneous rejection/criticism of 2% price inflation targeting, yet a defense of 3.5% NGDP as ‘just right’ which is itself dependent on the 2%. Site owner is talking about 3.5% only because of 2%. This is how the dialectic synthesis of ‘sublimation’ logic reveals itself in inconsistent discourse. 9% is ‘too high’ because it should be close to 2%, so 3.5% which is simultaneously both negated/rejected and also depended upon.

    It’s how ultimately the Marxist logic is pushed. NeoMarxism is a dialectic synthesis of Marxism and Post Modernism. It’s how escape hatches are established to continue pushing the same corrupt inner logic. NeoMarxists defend their views as ‘not Marxist’ because it negates Marxist concepts, while promoting the same Marxist logic of inherent conflict between people based on ‘group’ collectivist logic.

    The whole ‘NGDP level targeting’ is a dialectical synthesis of ‘negated’ price level targeting that also logically depends on it and ‘lifts it up’, an ‘auhfebung’. If instead price inflation in the US was 8% as ‘trend’, then the dependent NGDP level target synthesis would define 9.5% NGDP as ‘just right’ and 3.5% wouldn’t even enter the discourse.

    This pattern of pushing a logic by introducing dual inconsistent concepts, what looks like ‘negating’ something, is in fact carrying it to the future by synthesizing it with more subjectivity to cover up that dependency.

    As an example, the loudest ‘anti-racists’ who are convinced the world primarily operates according to racist motivations, the Critical Race Theorists, are themselves racists invoking racist conceptualizations that are unsolicited advanced as being ‘negated’ with them being the authority on negating their own negations using up other people along the way, but their ‘negations’ ARE THEMSELVES INVOCATIONS OF RACISM AND CARRIED INTO THE FUTURE as allegedly being observed when it’s really a psychological projection of their own dialectic set of ‘negations’ racist worldview.

    Beware the LOUD critics of price level targeting. For they are trying to sneak the logic of it in as dialectically synthesized to appear as if it is ‘correcting’ an inconsistency it sees outside its logic when in reality it is as fully dependent on the same logic it claims to be ‘negating’.

  2. Gravatar of Alex S. Alex S.
    18. November 2022 at 14:32

    Very convenient timing to post on this topic! I have two Q’s with some surrounding context.

    Since the expected path of NGDP E[NGDP] is a (the) key variable in the market monetarist framework it represents a sort of consensus forecast.

    Q1: What are the macro implications if the range of individual forecasters’ disagreements about the path of E[NGDP] increase?

    Divisia M4 is 34% higher than in December 2019 (note that this didn’t increase noticeably in 2008-2009 — it even decreased in that period and grew about 4% YoY thereafter in the 2010s). Currently, NGDP is only 19% higher than in December 2019. Now, M4 is growing at only 2% YoY the last three months or so. Thus, for inertial reasons money is very easy though marginally it appears to be tightening.

    So we have this level shift in the money supply that appears to be mostly permanent. We have a slowdown in the broader money supply growth rate. I see some unusually elevated degree of confluence of inertia from past policy actions and uncertainty about future policy actions.

    I suppose the key residual is what happens to money demand, which will be dependent on the E[NGDP] level path.

    Q2: I suppose my second question has to do with the limits of E[NGDP] as a sort of policy variable. If E[NGDP] is all that matters couldn’t the Fed increase the money supply by a factor of infinity and not cause any issue so long as it succeeds at keeping E[NGDP] constant.

    [Apologies if this became a long rambling post but this issue has been gnawing at me for some time and I’m not smart enough/motivated enough to untangle it myself.]

  3. Gravatar of George George
    18. November 2022 at 15:06



    All the Lefties that SHRIEKED IN GLEE as they watched the Hunter Biden laptop story banned on Twitter 2 years ago…

    …the Lefties that HOWLED IN DELIGHT as hundreds of thousands of Conservative accounts were suspended for trying to talk about/share laptop details…

    …the Lefties that chortled in mirth as their echo chamber was formed around them…


    Elon Musk fired almost everyone, the core team is down to just 50 people.

    And Twitter usage has reached at an all time high.

    If you do the math, it indicates the main purpose of the 7,500 Twitter employees was to censor/suppress ‘counter-revolutionary’, i.e. everything detrimental to the Democrat Party, narrative.

    I LOVE LOVE LOVE watching the authoritarian minded lefties have meltdowns because they can no longer depend on censorship, can no longer control the narrative, and who criticize Musk because the same radical left fake news msm is doing so. Hey, nobody wants to be smeared as a ‘conspiracy theorist’ by the conspiracy theory pushing msm do they? Conform or else! LOL



    grab popcorn.

  4. Gravatar of dtoh dtoh
    18. November 2022 at 15:25


    I’m not sure you can get ”back to the trend line after a steep decline caused by special factors” without generating inflation. It’s not like some kind of switch on a train track.

    To get back to the higher line, you need to temporarily increase NGDP growth, which I think pushes up both the real and nominal components of growth. If this is a good policy choice, why not choose the same policy (higher NDGP growth) after you get back to the trend line? There’s certainly no capital constraint. Is it a labor constraint? Is it a labor constraint when there are still 2 million people who haven’t rejoined the labor force after Covid?

  5. Gravatar of dtoh dtoh
    18. November 2022 at 15:27

    And one more thing. Please put a word count restriction on your two post rule. (I get tendinitis in my fingers from scrolling through the verbose posts.)

  6. Gravatar of Ricardo Ricardo
    18. November 2022 at 17:38

    I cannot wait for this upcoming recession/depression. I hope everything crashes. I hope the market drops 80%, so we can finally flush out all of these unprofitable companies operating on easy credit for the past ten years.

    I just hope the democrat, globalist, thugs don’t strike twice, and force the tax payer to foot the bill because Mr. Bank could borrow at zero percent lend at 1% then claim they are too big too fail. We need to let people lose their savings. They might finally learn the concept of risk. They might learn that some of these woke tech guru’s and financial institutions are charalatans holding a lot of bad “assets” ready to go bust once the interest rates increase and the easy money stops.

    We’ve got to punish these anti-scientific, anti-enlightenment, anti-free trading globalists who espouse their love for “free trade” and “free markets” in the form of “mercantilist agreements” and through their “monopoly on money”, and through forcing tax payers to pay for their bad and risk investments. They are a disease. They are a threat to humanity.

    Maybe the best way is through a fixed tax and fixed supply of currency. This would force congress to live within a budget, limiting the size of the federal government, while leaving room for real emergencies (say overiding the the fixed tax and fixed supply with 2/3rds majoirity.

    Because real emergencies don’t mean choosing sides over Kiev vs Donbass political issues, or going on religious crusades to change North Korean leadership. That’s not our problem.

  7. Gravatar of Rodrigo Rodrigo
    18. November 2022 at 18:15


    5 year breakevens are at 2.28% meaning the market believes the Fed will do enough to get back 2% inflation in the near future. Granted this would mean we don’t get back to previous ngdp growth path but adopt a new path, assuming similar ngdp growth levels. Similar to how (on the flip side) we never got back to our growth path post 2008, which lead to a slow and painful recovery. But What would be the repercussions of not going back down to the previous trend line? It feels as if the downside of both scenarios are not symmetrical as going back down would cause unnecessary pain while going back up post 2008 would have avoided it. I am sure there is a reason you would rather us go back to trend instead of adopting this new higher but still parallel, growth path. Is it you don’t believe the fed will be able to control inflation if we do adopt this new path? Or that the fed will do too much in trying to respond? At this point I am starting to think the recession everyone predicted in 2023 wont happen as, like you said, the fed has not tightened much.

  8. Gravatar of Ray Lopez Ray Lopez
    18. November 2022 at 20:59

    SS: “I understand that concepts such as easy and tight money are subjective” … wow, do I see a chink in Scott’s armour? The beginnings of wisdom? Given nobody seems to understand our man in LA, thus making his musings border on the metaphysical? Nah, I think it’s just a typical SS rhetorical throwaway phrase. Carry on, nothing here.

  9. Gravatar of Michael Sandifer Michael Sandifer
    18. November 2022 at 21:14

    But, why should we focus on past NGDP, even if recent, instead of forward-looking indicators? Currently, the 5 year breakeven is a bit below 2% in PCE terms.

    Also, we don’t have to drop NGDP back to the previous trend line. We could just sufficiently slow it to get it back on trend (let the trend catch up to the growth path, so to speak).

    And, the Fed has an asymmetric inflation target with an upside bias, so that rule does not mean the Fed promises to return to a lower previous growth path. It obviously does not imply NGDP level targeting, though the Fed could easily meet its mandate that way. It merely focuses on average inflation rates and unemployment. Level targeting is only discussed by this Fed when monetary policy has been tight. I agree that it’s not a good regime, but that’s what it is.

    Finally, I don’t trust that the previous NGDP growth path was optimal. Yes, in the long run it doesn’t matter, but we all know what Keynes said about the long run. I don’t think there’s such a thing as SRAS and SRAD. It’s more like medium-run AS and AD.

  10. Gravatar of bill bill
    19. November 2022 at 05:29

    10 year inflation breakeven is at 2.25% CPI (~2.0% PCE). 5 year BE is at 2.28%. Implied is those numbers is a PCE a teeny bit below 2% in years 2 thru 10.

  11. Gravatar of Spencer Spencer
    19. November 2022 at 06:38

    I don’t know why someone should be confused about the level of N-gDp?

    Sumner is correct. There’s historical evidence. The lags for R-gDp and N-gDp are interconnected. The lag in the rate-of-change in short-term money flows, the rhythmical oscillations, translate into the lag effect of long-term money flows, and vice versa. It’s just math.

  12. Gravatar of Spencer Spencer
    19. November 2022 at 06:49

    2022 is a good example. As long-term money flows have receded, short-term money flows have rebounded. Atlanta’s gDpnow forecast is @ 4.2% and Cleveland’s CPI inflation forecast for the 4th qtr. is @ 5.23%.

    Not that those numbers represent precise estimates, but they demonstrate the trend, the flows.

    And Barnett’s Divisia M4 is conterminous, not predictive. Barnett doesn’t use lags. That’s why I beat his estimates in early 1980 and in 1981.

  13. Gravatar of ssumner ssumner
    19. November 2022 at 09:28

    Alex, I suspect that the average NGDP forecast among business people and investors is more important than the distribution of forecasts, but I cannot be certain. I don’t view the monetary aggregate data as having much predictive value.

    On your second question, obviously at some point before infinity it would have an impact. But if you look at cases like Japan, where the monetary base growth has been enormous with very little NGDP growth, it does seem like NGDP is much more informative.

    dtoh, You said:

    “I’m not sure you can get ”back to the trend line after a steep decline caused by special factors” without generating inflation.”

    Yes, if there is a reduction in aggregate supply. But in that case the inflation would be a feature, not a bug.

    I’m not sure why you’d want to increase NGDP growth after getting back to the trend line. The whole point of a policy rule is to stick to it. Otherwise you end up with discretion. Remember the 1960s and 1970s?

    Rodrigo, You misunderstood my argument. I said they should have adopted level targeting in 2021. If they’d done so, we wouldn’t be in this situation—the overshoot would have been far smaller. I agree that getting back to the trend line would now be a mistake. But I don’t agree that we will continue from here on out at 3.5% NGDP growth. I predict much higher NGDP growth in Q4.

    The past year is why targeting TIPS spreads is not enough. A year ago the TIPS market predicted modest inflation. That’s why you need level targeting, or at least AIT.

    Michael, You said:

    “But, why should we focus on past NGDP, even if recent, instead of forward-looking indicators?”

    We look at the past to figure out what the Fed did wrong, to learn from mistakes, so that they can develop a policy regime that will prevent those mistakes in the future. That regime is level targeting.

    “we all know what Keynes said about the long run.”

    Of course Keynes was wrong; the long run matters a lot for what’s happening right now.

  14. Gravatar of dtoh dtoh
    19. November 2022 at 11:57


    “Yes, if there is a reduction in aggregate supply. But in that case the inflation would be a feature, not a bug.”

    Economically yes I agree. Politically maybe not.

    “The whole point of a policy rule is to stick to it.”

    I agree on that too, but if you set policy at a higher NGDP rate and it gets you a little more real growth and a little more inflation, isn’t that a good thing economically.

    Or is it the case that if you set the rate higher than the optimal policy rate, then any additional growth will be 100% inflation and you wont’ get any additional real growth?

  15. Gravatar of Travis Allison Travis Allison
    19. November 2022 at 16:48

    What do you mean exactly by, “Nominal wage growth has developed a lot of upward momentum, making it much harder to control.”?

    Do you mean that people have expectations for wages to increase by a lot and that expectation will be difficult to change? I suppose I am curious what you mean by “momentum”.

  16. Gravatar of ssumner ssumner
    19. November 2022 at 17:55

    dtoh, You said:

    “gets you a little more real growth and a little more inflation, isn’t that a good thing economically.”

    No, it’s really bad, as the long run effects are very negative. A shot of heroin also feels good in the short run.

    Travis, It’s partly a question of expectations, and partly a question of sticky long-term nominal wage contracts.

  17. Gravatar of dtoh dtoh
    19. November 2022 at 18:32

    I’m not talking about a one time shot of higher NGDP growth. I’m talking about a higher permanent target rate.

  18. Gravatar of Michael Sandifer Michael Sandifer
    19. November 2022 at 18:57


    You perhaps misunderstand my point about the Keynes quote. The point was the there haven’t been short-run AS and AD in the US in my lifetime.

    The Fed has been choking real growth for much of the last 4 decades. For example, core PCE inflation was well below the Fed’s implicit 2% target from 1996 until the recovery from the early 2000s recession, despite a sharp rise in oil prices and an RGDP boom. Sure, from an NGDP level targeting perspective, one can argue the Fed was right to restrain NGDP growth. However, that ignores the question of what the proper target should have been, and of course, the Fed wasn’t targeting NGDP. It was inflation targeting, and it’s not as if there was a long, stable period of NGDP growth prior. The 1999 to 2001 slowdown and recession were entirely unnecessary, with the Fed strangling many baby tech companies in their cribs.

    For another example, there’s the Great Recession and recovery. There’s not much disagreement here about whether there was tight money, but how tight and for how long?

    Even in 2018, when real growth exceeded 3% some quarters and core PCE was slightly above 2%, the 5 year inflation breakeven was well below 2% in core PCE terms. While Trump, whom I can’t stand, hurt himself with his stupid trade wars, the Fed compounded the problem with tight money. You call 2019 a “boom” year, but core PCE and the 5 year breakeven were well below 2% that year, with stock prices soaring. I take that last fact as evidence the economy was in catch-up growth mode after the Fed loosened policy, realizing its mistake.

  19. Gravatar of Michael Sandifer Michael Sandifer
    19. November 2022 at 20:34


    I think you have a point about wanting some more real growth and inflation. If we take the SRAS/SRAD model seriously, we should expect both RGDP growth and inflation to rise as the economy nears full employment. Under the Fed’s old 2% inflation targeting regime, that is not allowed, of course, so we never get a full recovery.

  20. Gravatar of dtoh dtoh
    20. November 2022 at 06:17

    Michael Sandifer,

    I agree. Two points.

    First, I think Scott will (he has in the past) argue that if you set the target above some “optimal” rate, then 100% of the additional growth is nominal and you get no additional real growth. I don’t believe that to be correct.

    Second, when you talk about full employment, it’s perhaps more important to look at the LFPR than the to look at the unemployment number. IMHO, the labor supply is a lot more elastic than the conventional wisdom which suggests that the unemployment rate dictates some kind of hard cap on the labor supply.

  21. Gravatar of Michael Sandifer Michael Sandifer
    20. November 2022 at 06:54


    I do suspect that tight money has been one factor, among others, in the declining prime age LFPR. I haven’t studied this question in much detail yet, but the demographics fit those who are hurt most by the tight money.

    There’s been talk of increasingly slow recoveries since the 1990 recession. It’s due to tighter money. We don’t even get full recoveries, because the SRAS/AD is really MRAS/AD. “M” is for “medium-run”.

  22. Gravatar of ssumner ssumner
    20. November 2022 at 10:35

    When you raise the trend rate of NGDP growth, you raise the effective tax rate on capital income, which (above some point) slows RGDP growth in the long run. I’d say the optimal NGDP rate is somewhere around 3% per capita, but whatever it is we should not change it in response to a policy mess-up, that would destroy the credibility of future policy.

  23. Gravatar of Michael Sandifer Michael Sandifer
    20. November 2022 at 14:52


    Of course we’d rather have 2% mean long-run inflation rather than 3%. And 0% would be better still, so why do you favor having any long-run mean inflation at all? Presumably, because you don’t want to cause too much pain in the shorter-run.

    The real question is, what NGDP level target do we choose to avoid too much unnecessary pain in what I call the medium-term? You’re making assumptions about productivity growth into the foreseeable future. Some others think productivity growth will be higher. Why not set it at a level that fits market forecasts for future growth at the time of adoption?

    In 2019, for example, if we take the stock market seriously, as a proxy for the second derivative of the expected NGDP path, the mean expected NGDP growth rate increased to 5.15% by the end of that year, with a 5 year inflation breakeven at just 1.68%. Realized mean NGDP growth was ~4.13%, and the S&P 500 rose 29% that year.

    If the economy was near equlibrium, as in 2019, as you’ve claimed, why the low realized and expected inflation numbers and why the 29% increase in stock prices? What mathematical model suggests those numbers are consistent with macro equilibrium?

  24. Gravatar of Michael Sandifer Michael Sandifer
    20. November 2022 at 15:02

    I should mention, that the imputed mean expected NGDP growth rate mentioned above assumes that the 2019 NGDP growth rate, which was close to the mean of the prior 10 years, was the baseline expected growth rate. That’s just for convenience. With analysis in further depth, there are reasons to believe the baseline expected growth rate was actually higher.

  25. Gravatar of dtoh dtoh
    20. November 2022 at 23:43

    Just to be clear nobody is suggesting raising the rate in response to a policy mess up. I totally agree that’s a bad idea.

    The “inflation tax” issue is orthogonal to the question I’m raising. That problem can be solved by indexing gains to inflation. Or mostly solved just by lowering the capital gains tax rate.

    The question that remains and the question I’m trying to ask is what is the optimal long range NGDP target rate. You say you think 3% is about right. But if 4% will get you some additional real growth, why wouldn’t (ignoring the inflation tax issue) a 4% target be better than a 3% target rate.

  26. Gravatar of madm madm
    21. November 2022 at 07:48

    I tend to agree but I do think you need to account for forward guidance, the impact of business and household leverage, and long and variable lags on the Fed’s thinking about the appropriate response.

    Let me start by oversimplifying with the Fisher equation.

    In the second half of 2019 we had full employment, fed funds rate at 1.5-2.5%, and core inflation on target around 2%. That’s a real rate of around 0%.

    In 2021, we were still well below full employment (arguably the gap was close to closed by Q3), fed funds rate was 0% and core inflation averaged 3.6%. That’s a real rate of around -3.6%.

    In 2022, we are roughly back to full employment, fed funds rate has averaged 1.5 and core inflation running at 7%. That’s a negative real rate of -5.5%, which at full employment seems completely unjustifiable.

    This month we still have full employment and finally have core inflation down to around 3.5% for October, which if it holds there or below again this month will finally give us a positive real rate now that the fed funds is above 3.75%. That said, we are still very vulnerable to another inflation shock from the energy market or more trade warring with and/or lockdowns from China.

    Despite what looks like a way too loose a policy response, the Fed was able to get away with a policy rate that was so low by clearly signaling rates would be going up. Their argument has been that this is a better policy than jacking up rates immediately because it gives them more time to judge the demand response–the long and variable lags argument. The pace of hikes was clearly too slow ex post for at least half of the year but it significantly quickened since then and the guidance got a lot sharper. Judging by forward rates, investors now seem to think we’re already pretty close to the terminal rate. And that will likely line up with the FOMC’s next set of projections. That’s all baked into the yield curve, so averaging across the curve we have slightly higher real rates than implied by looking at the overnight rate.

    What’s puzzling is that investors’ implied odds of inflation above 3% for the next 5 years are still almost even money (source: Minn Fed), yet the market-implied average inflation has inflation back on target within a year (source: ICE). At the same time investors think nominal rates are not going to move much above 5% and will then decline after about a year consistent with Fed guidance (source: CME). That contradictory mix of signals suggests investors are probably putting a lot of weight on recession. How can a barely positive real rate–even with forward guidance–be contractionary at all when the lesson of the Taylor rule was that rates needed to move much more than 1:1 with inflation? Those households/businesses who are heavily indebted in proportion to their incomes/EBITDA, which was much less prevalent during 70-80s era of inflation, will feel a disproportionate pinch from higher rates. Yet to be seen whether their curtailed demand will be made up from less leveraged businesses and households but historically it doesn’t look good.

  27. Gravatar of ssumner ssumner
    21. November 2022 at 08:57

    Michael , You said:

    I should mention, that the imputed mean expected NGDP growth rate mentioned above assumes that the 2019 NGDP growth rate, which was close to the mean of the prior 10 years, was the baseline expected growth rate.I should mention, that the imputed mean expected NGDP growth rate mentioned above assumes that the 2019 NGDP growth rate, which was close to the mean of the prior 10 years, was the baseline expected growth rate.”

    Unemployment fell from 10% to 3.5% over that period. So you are saying that markets expected unemployment to fall to negative 6.5% over the next decade?

    dtoh, I’m not sure that 4% gets you more growth than 3%, and I see almost no chance the capital income tax system will be indexed, which would be an administrative nightmare.

    madm, Interesting comments, but I have doubts about the “long and variable lags”. See what you think of my new Econlog post:


    Also, be careful with ex post real interest rates, as the ex ante rates may differ.

  28. Gravatar of Michael Sandifer Michael Sandifer
    21. November 2022 at 10:18


    You replied:

    “Unemployment fell from 10% to 3.5% over that period. So you are saying that markets expected unemployment to fall to negative 6.5% over the next decade?”

    No, obviously unemployment would bottom out somewhere in the positive range at a sustainable higher real growth rate, implying increasing productivity growth. That’s after some measure of catch up growth.

    I wouldn’t say we were terribly far from full-employment in 2019. Afterall, how could we be? But, it also doesn’t mean we were as close as many seemed to assume. I recall so many economists, including those at the Fed, for years assumed the NAIRU was around 5-to-5.5%. Obviously, they were way off, though granted, it’s a fluctuating variable.

    What’s your explanation for high stock price appreciation during years where you think we’re very close to equilibrium, with well-below target inflation expectations?

  29. Gravatar of Ram Ram
    21. November 2022 at 11:59

    It looks like the market expects PCEPI inflation to run at about 1.8% on average for the next 5 years, which is a bit less than the desired long-run rate of 2%. This is the same market that expects the Fed to raise the FFR by another percentage point in the next few months and then stop increasing rates and perhaps gradually begin lowering them. That suggests that the market thinks we’re getting close to a policy rate consistent with (very slow) reversion to the desired price level path. If that’s right, then they may not need to raise rates too much more than that to get a somewhat faster pace of reversion, maybe 2 percentage points. As much as I grant that the market is smarter than me, something tells me it’s being overly optimistic here. If I were a betting man, I would bet on higher than 1.8% inflation over the next five years conditional on not much too much further movement in the policy rate.

  30. Gravatar of dtoh dtoh
    21. November 2022 at 19:44


    You said,

    “I’m not sure that 4% gets you more growth than 3%, and I see almost no chance the capital income tax system will be indexed, which would be an administrative nightmare.”

    Why do question whether 4% gets you more growth? What is the constraint?

    Regarding taxes, I agree indexing would be an administrative PITA, but as I said simply lowering the capital gains tax rate mostly solves the inflation problem

  31. Gravatar of madm madm
    22. November 2022 at 07:05

    Thanks Scott. Re “I have doubts about the “long and variable lags”. See what you think of my new Econlog post”

    We’re broadly on the same page here. When I think about lags, I exclusively think in terms of your bullet point 2. But I also include unanticipated shifts in forward guidance as being part of the policy arsenal, and I would think you would agree with that. I agree that there aren’t many lags in the market response to policy news but spending and investment responses can be slower moving for a lot of reasons.

    In the first six months of 2022, even with its forward guidance, the Fed was indefensibly slow. I think the employment and inflation data made the case for lifting rates as early as September 2021. But certainly by Jan of this year, the target rate should have been lifted at least 50 bps, followed by another 100 bps in March. The Fed may have been able to slow things down or even pause at that point to assess your category 2 lags. Given the post-pandemic asset values propped up by low rates, historically high levels of aggregate leverage and the more skewed distribution of leverage in the economy, it probably takes a lot less in terms of rate hikes to slow the economy. I know you’re not a fan of credit-linked explanations Fed policy, but I think this is a pretty standard result in a financial accelerator model. All of which is to say, R* is probably lower than in the past. I do think the R* framework is a massive oversimplification of a complex economic system, which I’d imagine you’d agree with given your preference for nominal GDP targeting.

    The consequences of the Fed missing the optimal rate path are interesting. I’m definitely not in the reflexive hawk camp where being 1% under target for a couple of quarters is somehow incredibly destructive. But it definitely seems right that it becomes more costly to unwind policy misses the longer they go uncorrected. On that front, the bright side of higher rates has been the collapse of the many pie-in-the-sky and/or fraudulent crypto schemes. Would-be money substitutes look like a crappy store of value when money market funds are paying positive nominal rates.

    Re “Also, be careful with ex post real interest rates, as the ex ante rates may differ.”

    Sure, I was cutting some corners there but given FF is an overnight rate and the choice of inflation expectations is between the most recent inflation report and the 1-year inflation swap rate, I’m going with the inflation report to calculate the real rate, especially when the market-based inflation rates were all signaling expected inflation that was persistently lower than realized over this period. Was only trying to make the superficial case that a Taylor rule would tell you policy was way too loose. But once you move to a more forward looking measures that embed the Fed’s forward guidance that looseness is much less clear cut, especially now.

    When I wrote my first comment, I forgot that Cleveland Fed has market implied inflation data by tenor that can shed some light on this:

    The highest the 1-year inflation breakeven got in 2022 was 4.25% in June, and it has only averaged 3.3% this year (source: Cleveland Fed). The implied real 1-year Treasury rate was at its lowest at -4% in June and averaged -2%. As of Nov, the 1yr breakeven is 3.2% and the 1-year real rate is barely positive at 0.4%.

    Going further out on the yield curve though, the 10-year breakeven has been pretty stable and at it highest in November, just below 2.5%, and has averaged 2.3% over the year (just below target on core PCE basis). But the lowest real 10-year yield was 0.4% back in March and it has averaged 1% for the year. As of Nov, the real 10-year was 1.9%. So it depends on how you average across the short- and long-term rates and where you stand on R* to decide how tight things are but the case that policy was loose was certainly much stronger in the first half of the year. I’ll caveat that it’s harder to interpret market implied rates the further out you go on the curve because of well-documented term premium distortions and liquidity issues in the underlying markets. I’m also not convinced that the markets have this stuff as well figured out as we economists would often like to think.

    Btw, it’s been quite a while since I’ve posted a comment here. I interacted with you in the comments during the height of the financial crisis when we were all trying to scramble to figure out macro policy in real time. I’m impressed that you’ve kept the blog going all this time and still respond thoughtfully and openly to comments. Your voice remains unique, opinionated, thoughtful, and, best of all, not blinded by partisan loyalties. Thanks for that!

  32. Gravatar of ssumner ssumner
    23. November 2022 at 09:02

    Michael, Stocks did well for many reasons, a good economy, more pricing power, and low long-term real rates among them.

    Ram, You said:

    “If I were a betting man, I would bet on higher than 1.8% inflation over the next five years conditional on not much too much further movement in the policy rate.”

    So would I. BTW, I heard the same arguments a year ago, and they were wrong.

    dtoh, The PPF is the constraint. We are talking about long run monetary policy, a period over which money is neutral. Do you think Bangladesh could get rich by printing money?

    madm, Very thoughtful comment–I mostly agree. A few comments:

    You said:

    “I agree that there aren’t many lags in the market response to policy news but spending and investment responses can be slower moving for a lot of reasons.”

    In the 1930s, stocks were far more volatile than usual, as was monetary policy. And yet industrial production moved almost one for one with stock prices. If monetary policy was driving stocks during that period (as I strongly believe) then there must have been a very short lag on output.

    You said: “But it definitely seems right that it becomes more costly to unwind policy misses the longer they go uncorrected.”

    I’d add that it’s more costly to unwind errors when you don’t have a credible level targeting regime. Level targeting tends to preemptively move longer term rates in the appropriate direction, and also tends to somewhat anchor nominal wage rates.

  33. Gravatar of dtoh dtoh
    23. November 2022 at 14:56

    “The PPF is the constraint.”

    Yes but my question is how do we know that a 3% NGDP target gets you to the PPF.

  34. Gravatar of Michael Sandifer Michael Sandifer
    25. November 2022 at 10:59


    You replied:

    “Michael, Stocks did well for many reasons, a good economy, more pricing power, and low long-term real rates among them.”

    A “good economy” in equilibrium should not cause stock prices to rise. The key words in that sentence are “in equilibrium”, as you’ve said you thought the economy was very close to equlibrium in 2019. Economic growth expectations and/or earnings expectations would have increase to cause stock prices to rise 29% while in macro equilibrium. Did pricing power suddenly increase that much in 2019?

    Long-term real rates did fall nearly a basis point in 2019 as the yield curve went negative, but the most likely explanation for that would seem to be anticipated Fed rate cuts which began in August 2019.

    Do you really think the factors you cite explain stock market performance in 2019?

  35. Gravatar of dtoh dtoh
    25. November 2022 at 22:53

    Michael Sandifer,
    Assuming EMH and all that stuff, if rates are low and drop, it will translate into a very big price movement in the underlying asset.

  36. Gravatar of Michael Sandifer Michael Sandifer
    26. November 2022 at 10:45


    That’s simply untrue. Look at the case of Japan. Their interest rates fell with their stock prices, because their economic growth prospects faded, primarily due to demographics.

  37. Gravatar of dtoh dtoh
    26. November 2022 at 13:10

    Michael Sandifer,
    That’s exactly what I’m saying. Current asset price equals future earnings/cash stream discounted by the risk adjusted interest rate. If expected earnings decrease, the effect on the asset price will be in the opposite of direction of a change in the discount rate. If there is no change in expected earnings, small changes in interest rates will cause large increases in the asset price.

    Also Japan was a bit of weird case because of messed up policy. Life insurers who were big equity holders were not allowed to pay policy dividends out of capital gains which resulted in an artificially inflated stock market. After GS and CSFB figured out a way to allow the life companies to convert their gains into interest income by selling massive amounts of put options on their equity holdings, the Japanese market crashed.

  38. Gravatar of Michael Sandifer Michael Sandifer
    27. November 2022 at 06:23


    You seem to be referencing the standard finance model for the macro determination of stock prices, which makes sense. Almost anyone would. However, I think that model is wrong.

    It may sound silly, but I don’t think real interest rates matter at all for stock prices. Real rates and stock prices are both dependent variables. What matters is expected NGDP growth.

    What many economists fail to realize, as I’ve confirmed many times personally, is that the longer-run mean S&P 500 earnings yield and mean NGDP growth rates are basically equal in the US. For example, since 1962, the difference is only about 0.2%. This supports the view that changes in the S&P 500, for example, are a good proxy for changes in expected NGDP growth. It’s another way of saying that in an equation to model the “price” of the US economy, in the longer-run, the discount rate would be the same as that for the S&P 500. Earnings are more volatile than NGDP, but otherwise stock prices are a good indicator of expected economic growth.

    So, one way to model it is:

    present earnings(1+g)^t/G = Price in “t” years into the future, where “g” is the mean expected earnings growth rate and “G” is the current NGDP growth rate. “g” in this case would also be a good proxy for expected “G”.

  39. Gravatar of ssumner ssumner
    27. November 2022 at 07:56

    dtoh, We don’t know anything for certain, but there’s no reason to suppose a 4% target would get keep you on the PPF any more effectively than a 3% target.

    Michael, Yes, I think those three factors explain much of the increase, although other factors undoubtedly played a role.

  40. Gravatar of dtoh dtoh
    27. November 2022 at 17:49

    Scott, Tell me if I am interpreting your response incorrectly.

    You believe raising the target rate is unlikely to get you any additional real growth. I.e. you think any increase above 3% in the NGDP target will translate 100% into nominal growth and 0% real growth.

    Given how beneficial real growth is, your reluctance to even consider a higher target suggests you are highly confident that a higher NGDP target will not produce any additional real growth.

    Is that a correct interpretation?

    Parallel question – If the NGDP target is lowered below 3%, will the long term reduction in growth be a) 100% nominal, b) 100% real, or c) some combination of real and nominal?

  41. Gravatar of dtoh dtoh
    27. November 2022 at 18:01

    Michael Sandifer,

    I have seen your comments about this before and generally agree. From a logical perspective, it passes the sniff test. Essentially, what you’re saying is that there is a close correlation between corporate earnings, real rates and NGDP growth.

    My questions would be: a) which way is causality pointing, and b) what is the impact of exogenous factors (e.g. tax rates on capital, increasing degrees of monopolization, etc.)

    I think you’re probably right in the long term, but in the short term, I think relative asset prices are very sensitive to changes in real after tax risk adjusted rates. Or to put it more accurately, I think asset prices react very quickly to changes in relative risk adjusted rates.

  42. Gravatar of Michael Sandifer Michael Sandifer
    27. November 2022 at 18:59


    There are certainly companies in the S&P 500 with considerable pricing power. Apple comes to mind, for example. But, I don’t see why or how that’d be a major factor to consider in 2019, in particular.

    Sure, the drop in real rates alone could explain much of the stock price increase in 2019, given the standard finance model, but I don’t think that model is sound. Why did real rates fall? That’s the important question. You’ve acknowledged that lower real rates do not always lead to lower stock prices.

  43. Gravatar of dtoh dtoh
    27. November 2022 at 23:32

    Michael, Scott,
    I think you may be making this unnecessarily complicated by thinking of rates and prices as separate variables. For some assets, we regularly use multiple terms to express the same thing. Price and yield in the case of bonds. Volatility and price in the case of options. Yields and price for commercial real estate. Etc. In the case of equities, it’s less convenient to do this, but fundamentally the equivalence still exists, and if we wanted we could just as easily express equity prices as earnings yields.

  44. Gravatar of Michael Sandifer Michael Sandifer
    28. November 2022 at 05:41


    Yes, your comments seem to indicate you get the gist of it.

    To answer your questions:

    a.) NGDP growth expectations determine stock prices on the macro level. I think there’s a certain rate of expected NGDP growth each year, for example, and then real and nominal NGDP shocks lead to changes in stock prices. However, for temporary real shocks the influence on stock prices will be minimal if the Fed keeps expected aggregate demand stable. Hence, almost all relatively large stock market price movements are caused by the Fed.

    b.) Monopolization is a gradual process at the macro level, so I don’t see it causing abrupt changes to stock prices. Changes in tax policy can obviously cause real shocks with lasting effects on prices. However, I don’t think bracket creep, which has been greatly reduced since the 1970s has much effect on stock prices when inflation is low.

    I think real interest rates are a mostly a distraction, whether one is talking about the stance of monetary policy, or factors driving changes in stock prices. Tighter money can lead to lower real rates in the short-term, or higher real rates, depending on the circumstances. It lead to lower real rates after the Great Recession, for example, and higher real rates during the current tightening cycle. In the long-run, real rates are determined by real factors, of course.

    I want to emphasize something in this thread that I’ve only implied. The discount rate for the S&P 500 spot price is the earnings yield. In my discussions with economists and finance-types, many have been surprised by that fact, as it seems that they don’t spend much time explicitly playing with the relevant equations.

  45. Gravatar of dtoh dtoh
    28. November 2022 at 07:54

    I would think of it this way…. the earnings yield and the stock price are exactly the same thing. They are just two different ways of measuring or saying the same thing. We could just as easily quote stock prices as an earnings yield. I come from a fixed income and options background so this is obvious and innate to me, but I think it’s a less obvious way of thinking about things if you come at it from an equities perspective.

  46. Gravatar of Michael Sandifer Michael Sandifer
    28. November 2022 at 16:42


    If you think the earnings yield and the stock price are the same thing, then it supports my view, which is apparently what you meant with your earlier reply. I suspect you mean that the reciprocal of the earnings yield, the P/E ratio and stock price are the same thing. They aren’t, of course, because the earnings yield includes price, but the P/E multiple is an essential feature. And, of course, the P/E is highly correlated with price.

  47. Gravatar of Michael Sandifer Michael Sandifer
    29. November 2022 at 05:35


    If you look at countries in western Europe, in addition to Japan, you find mean earnings yields on broad stock indexes that are higher than the mean NGDP growth rates, depsite low long-term real interest rates. That makes sense, given the relatively worse demographic prospects, ceteris paribus.

    But, if you’re right about the US having similarly declining real growth prospects, albeit not as bad, why isn’t the mean earnings yield running above the mean NGDP growth rate? Do you think this is a trend yet to come, or do you think factors like pricing power are making the difference? I assume you don’t think changes in pricing power occur within, say, single years to cause stock prices to rise as much as 29%. That would seem silly. Increasing pricing power would be more of a long-run phenomenon, as companies like Apple and Tesla mature.

  48. Gravatar of dtoh dtoh
    29. November 2022 at 15:37

    Earnings yield and price express exactly the same thing. They are mathematically equivalent. At any point in time earnings are a constant. For any given earnings yield, there is only one price, and for any given price there is only one earnings yield. If equities were traded on the basis of yield instead of price, you would then be arguing that price is different because it includes earnings yield.

    But… it is a mistake to think that they are different or that there is any kind of causal connection between the two. It would be like debating the causal connection between the PE ratio and the earnings yield. They are just different ways of expressing the same thing. Whether people talk about price or about yield is just a convention based on convenience and historical practice.

  49. Gravatar of Michael Sandifer Michael Sandifer
    30. November 2022 at 06:39


    Sure, I think about stocks in terms of earnings yield, but it’s not actually equivalent to bond yields, for example, if we’re talking about the typical quote using trailing 12 month earnings.

    I impute the forward P/E in market prices when thinking about my approach. That’s different than what you seem to be talking about. I’m not even sure what point you’re trying to make or what the relevance is.

Leave a Reply