Archive
The Phillips Curve is Still Working Just Fine
About five and a half years ago, I wrote a blog article entitled “The Phillips Curve is Working Just Fine, Thanks”, in response to the exhaustively-repeated nonsense that the ‘Phillips Curve is Broken.’ This nonsense never really goes away, but last week Fed Governor Waller delivered a speech on “The Unstable Phillips Curve,” derived from the same nonsense, and I felt duty-bound to resurrect my prior article and update it. The Phillips Curve has not been unstable at all, over the last quarter century at least. Here is my original article, linked here:

I must say that it is discouraging how often I have to write about the Phillips Curve.
The Phillips Curve is a very simple idea and a very powerful model. It simply says that when labor is in short supply, its price goes up. In other words: labor, like everything else, is traded in the context of supply and demand, and the price is sensitive to the balance of supply and demand.
Somewhere along the line, people decided that what Phillips really meant was that low unemployment caused consumer price inflation. It turns out that doesn’t really work (see chart, source BLS, showing unemployment versus CPI since 1997).
Accordingly, since the Phillips Curve is “broken,” lots of work has been done to resurrect it by “augmenting” it with expectations. This also does not work, although if you add enough variables to any model you will eventually get a decent fit.
And so here we are, with Federal Reserve officials and blue-chip economists alike bemoaning that the Fed has “only one model, and it’s broken,” when it never really worked in the first place. (Incidentally, the monetary model that relates money and velocity (via interest rates) to the price level works quite well, but apparently they haven’t gotten around to rediscovering monetarism at the Fed).
But the problem is not in our stars, but in ourselves. There is nothing wrong with the Phillips Curve. The title of William Phillips’ original paper is “The Relation between Unemployment and the Rate of Change of Money Wage Rates in the United Kingdom, 1861-1957.” Note that there is nothing in that title about consumer inflation! Here is the actual Phillips Curve in the US over the last 20 years, relating the Unemployment Rate to wages 9 months later.
The trendline here is a simple power function and actually resembles the shape of Phillips’ original curve. The R-squared of 0.91, I think, sufficiently rehabilitates Phillips. Don’t you?
I haven’t done anything tricky here. The Atlanta Fed Wage Growth Tracker is a relevant measure of wages which tracks the change in the wages of continuously-employed persons, and so avoids composition effects such as the fact that when unemployment drops, lower-quality workers (who earn lower wages) are the last to be hired. The 9-month lag is a reasonable response time for employers to respond to labor conditions when they are changing rapidly such as in 2009…but even with no lag, the R-squared is still 0.73 or so, despite the rapid changes in the Unemployment Rate in 2008-09.
So let Phillips rest in peace with his considerable contribution in place. Blame the lack of inflation on someone else.
Before I add to my rant, let me update the chart above with data since then, including the pandemic. The green dots in the chart below correspond to the dots in the chart above; the blue dots are for the period since then.
Amazingly, even during the pandemic and post-pandemic period, the Phillips Curve did a pretty decent job of describing the basic shape of this relationship. The dots overall are a bit higher; that’s attributable I think to the fact that inflation itself is higher and I’ve done this chart in nominal terms. There is some money illusion operating (or else the latest dots would be a lot higher), but it’s still a pretty nice fit, considering. I’ve preserved the prior regression line, but it doesn’t really shift very much.
In fact, the deviation prior to the pandemic – the little knot of blue dots to the left – are somewhat more surprising in a way, given the much lower economic volatility that there was when those points were laid down. But in any event, though, there is nothing obviously wrong with the Phillips Curve.
Now, it is true that the Unemployment Rate and the rate of consumer inflation have not been particularly well-behaved. But that isn’t a new phenomenon; that particular inconvenience has been that way for decades. The reason is pretty straightforward, and only confusing if you spent too much time getting a PhD and getting taught dumb things: the connection between wages and prices is not 1:1. It’s not constant. And there’s no particular reason that it should be, because labor is just one input into production costs, and the cost of production just affects the supply side of the supply/demand interplay which determines price. The really weird thing is that anyone ever thought that prices would be set by taking the current wage cost and adding a simple and stable markup.
A wage is just the price of labor, which is set in the market for labor, which involves the demand for labor and the supply of labor. The supply of labor changes very slowly. The demand for labor moves with the economic cycle. When the economic cycle is ebbing, the demand for labor falls – and that causes the quantity of labor demanded to decline (the unemployment rate goes up) as it also causes the price of labor to fall. That’s what happens when a demand curve shifts leftward on a mostly-static supply curve: Q down, P down. When the economic cycle is flowing, the demand for labor rises, which causes the quantity of labor demanded to increase (the unemployment rate declines) and the price of labor to rise. It isn’t that hard. In fact, you learn that in pretty much the first semester of economics.
It’s those later semesters that screw up economists, encouraging them to design complicated models that are very pretty but don’t necessarily relate to real-world dynamics. We should not be at all surprised when those models don’t work in the real world.
But don’t blame Phillips.
The Powell of Positive Thinking
Yes: Federal Reserve Chairman Powell was very hawkish at his Congressional testimony on Tuesday and Wednesday. He clearly signaled (again) that once Fed overnight policy rates reach a peak, they would not be declining for a while. He additionally signaled that the peak probably will be higher than previously signaled (I’ve been saying and thinking 5% for a while, but it’s going to be higher), and even signaled the increasing likelihood of a return to 50bp hikes after the recent deceleration to 25bps.
This latter point, in my view, is the least likely since all of the reasons for the step down to 25bps remain valid: whether the peak is 5% or 6%, it is relatively nearby and the confidence that we should have that rates have not risen enough should therefore be decreasing rapidly. Moreover, since monetary policy works with a lag and there has been very little lag since the aggressive tightening campaign began, it would be reasonable to slow down or stop to assess the effect that prior hikes have had.
But here is the bigger point, and one that Powell did not broach. There is really not much evidence at all that the Fed’s hikes to date have affected inflation. It is completely an article of faith that they surely will, but this is not the same as saying that they have. Consider for a moment: in what way could we plausibly argue that rate hikes so far have been responsible for the decline in inflation? The decline in inflation has been entirely from the goods sector, and a good portion of that has been from used cars returning to a normal level (meaning, in line with the growth in money) after having overshot. How exactly has monetary policy driven down the prices of goods?
This is not to say that higher interest rates have not affected economic activity, and this (to me) is the real surprise: given the amount of leverage extant in the corporate world, it amazes me that we haven’t seen a more-serious retrenchment. Some of this is pent-up demand that still needs to be satisfied, for example in housing where significant rate hikes would normally dampen housing demand substantially and seems to have. However, there is a severe shortage of housing in the country and so construction continues (and home prices, while they have fallen slightly, show no signs of the collapse that so many have forecast). Higher rates are also rippling through the commercial MBS market, as many commercial landlords have inexplicably financed their projects with floating rate debt and where the cost of leverage can make or break the project.
Higher interest rates, on the other hand, tend to support residential rents, at least until unemployment eventually rises appreciably. I think perhaps that not many economists are landlords, but higher costs tend to not result in a desire to charge lower rents. On the commercial side, leases are for longer and turnover is more costly, but the average residential landlord these days is not facing a shortage of demand.
So where have rate hikes caused inflation to decline? Judging from the fact that Median CPI just set a new high, I think the answer is pretty plain: they haven’t. And yet, the Fed believes that if they keep hiking, inflation will fall into place. Where else can we more plainly see at work the maxim that “if a piece doesn’t fit, you’re not using a big enough hammer?” Or maybe, this is just a reflection of the notion that if you want something bad enough, the wanting itself will cause the thing to happen. [N.B. this is really more in line with the prescription from Napoleon Hill’s classic book “Think and Grow Rich”, but the title of Peale’s equally-classic “The Power of Positive Thinking” suggested a catchier title for this article. Consider it poetic license.]
Moreover, what we have seen is that higher interest rates have had the predicted effect on money velocity. Although I have elsewhere noted that part of the rebound in money velocity so far is due to the ‘spring force’ effect, there is substantial evidence that one of the main drivers of money velocity is the interest rate earned on non-cash balances. Enough so, in fact, that I wrote about the connection in June 2022 in a piece entitled “The Coming Rise in Money Velocity,” before the recent surge in velocity began. [I’d also call your attention to a recently-published article by Samuel Reynard of the Swiss National Bank, “Central bank balance sheet, money, and inflation,” where he incorporates money velocity into his adjusted money supply growth figure. Reynard is one of the last monetarists extant in central banking circles.]
Now, nothing that I have just written is going to deter Powell & Co from continuing to hike rates until demand is finally crushed and, according to their faith but in the absence of evidence to date, inflation will decelerate back to where they want it. But with long-term inflation breakevens priced at levels mirroring that faith, it is worth questioning whether there is some value in being apostate.
Is Inflation Mean-Reverting?
Over the last couple of decades, the assumption that inflation is mean-reverting to something approximating the Fed’s target level (or to where inflation expectations are supposedly – without any evidence advanced to support the notion – ‘anchored’) has become a key component of most economists’ models. I’ve pointed out a number of times in podcasts (including my own Inflation Guy Podcast as well as numerous others) and in articles that after a quarter-century of having low and stable inflation any model which did not assume mean-reversion has been discarded because it made bad predictions over that period compared to one which did.
A critical follow-up question is whether a model should assume mean reversion in inflation. My observation implicitly says that it should not. If I’m wrong, and inflation in fact is mean-reverting, then the right models won and there’s no real problem.
So, did the right models win?
There are many sophisticated ways to test for mean reversion, but an intuitive one is this: for a given current level of inflation, which is a better guess: (a) inflation will be closer to the ‘mean’ in the next period; (b) inflation will be about the same distance from the mean (homeostasis), neither pulling towards the mean nor pushing away from the mean; or (c) inflation will be further away from the mean, such that deviations from mean get amplified over time. In case (b) we would say that inflation itself has momentum; in case (c) we would say the acceleration of inflation has momentum. The latter case seems an unlikely case of extreme instability: it says that once prices move away from equilibrium, the economy either enters into an inflationary spiral or a deflationary spiral with no clear end. While this clearly can eventually happen in the hyperinflation case, those cases seem to have other causes that tend to amplify the swings (notably, an accelerating loss of faith in the currency itself).
Let’s consider case (a) and (b), and look at some historical data.
The chart below shows the period 1957-2022. The x-axis indicates the current level of inflation, (I collapse the range from -0.5% to +0.5% and call it 0%, +0.5% to +1.5% and call it 1%, etc), and the y-axis shows the average inflation over the subsequent one year. So, the point that is at [2%, 2.3%] shows that between 1957 and 2022, if inflation was between 1.5% and 2.5% then the average inflation over the ensuing 12 months was 2.3%.
I’ve drawn a line that indicates inflation at the same level at the point of observation and subsequently (x=y). Notice that for any number below x=2%, y tends towards 2%. This shows that when the current reading is very low inflation or deflation, the subsequent year we tend to get something close to 2%. Notice that at higher rates of inflation, the dots are below the line – meaning that if inflation is high, the following year tends to see inflation closer to the target. So, this is what we would think mean reversion would look like (and FWIW, it is more pronounced if you choose a longer historical period but because the next chart I am showing is core CPI and we only have data to 1957, I wanted to use the same range).
Case closed! Inflation mean reverts!
Well, not exactly. This is headline inflation. We already know that food and energy tend to mean-revert; that is, after all, why economists exclude food and energy – because we know that high energy readings lead to high inflation prints, and we don’t want monetary policy to overreact to inflation that isn’t really persistent. So, let’s look at core instead.
This chart looks different in key aspects. Except for very high core readings (with comparatively few observations that happen to coincide with when Volcker was aggressively tightening policy), the best estimate for core inflation over the next 12 months is not something closer to the assumed mean; the best estimate is the same level as what we have right now.
What that means – and it is super important – is that inflation has momentum. Keep in mind that during most of the period shown here, the Federal Reserve was actively trying to make inflation mean-revert. And they didn’t succeed, at least on a one-year basis.
Well, monetary policy works with long and variable lags, right? How about core inflation over the period 12-24 months from now? Surely then we should see some mean reversion?
The answer, at least for core inflation, is decidedly no…except for very high current readings of inflation.
Two takeaways:
- Inflation has momentum. This means that forecasting core inflation to return to the target level, just because we think it should, is a bad forecasting approach.
- Monetary policy seems to have had, at least over this period, very little effect. Generously, it didn’t have effect on average…so perhaps sometimes the Fed overshot and other times it didn’t do enough. There is indeed a range. For example: starting from 5% y/y core inflation (between 4.5% and 5.5%), the 10th percentile of the 1y CPI outcomes after that was 3.5% and the 90th percentile was 6.0%. Starting from 7%, the 10th percentile was 3.1% and the 90th was 9.6%. So the average includes some times when inflation kept going up and some times when it was going back down.
The corollary to the second takeaway…call it takeaway 2a…is this: by the same token, there’s not a lot of reason for the Fed to be super aggressive raising rates to rein in inflation. We know that they can do harm. It’s less clear that they can do a whole lot of good!
Oil Be Home For Christmas
As a general rule, don’t trade on pre-holiday thin-liquidity sessions. There can be amazing-seeming opportunities, but price can still get shoved in your face by whoever it is who feels like pushing markets around.
A prime example today is the energy market, where front-month oil prices are down nearly 4% at this writing. Recently, energy futures have been regularly jammed lower overnight in low-liquidity conditions and then have recovered during the day. There is a structural shortage of energy globally at the moment, and inventories are low…but sentiment is also very poor and as I’ve shown before, open interest has been in a downtrend for years – aggregate open interest in NYMEX Crude hasn’t been lower since 2012.
So, it’s a market ripe for pushing around and the day before Thanksgiving is probably not the day to take a stand by getting long even when the reasons given for the selloff are nonsense. Today, the story is again about the price cap on Russian oil that is being implemented soon by the US and EU. Market participants seem to struggle with Econ 101 here. A price cap has one of two effects in the market under consideration: if the price cap is set above the market-clearing price, it has no effect. If the price cap is set below the market-clearing price, it leads to shortages as suppliers – in this case, Russia – won’t supply as much oil (if any) to the capped market when there are other uncapped markets (say, China and India). There is probably an area near the price cap where the cost of switching to delivering oil in other markets is higher than the gain from switching deliveries, but that’s only in round 1 of the game theoretic outcome.[1]
In this case, since only the price from one supplier is capped, the result should be higher prices in the markets than otherwise since once price exceeds the cap, one supplier is lost. The chart below shows the classic outcome. Below the cap, the supply curve is normal. Above the cap, the supply curve is left-shifted.
This leads, at least in a frictionless market (which this isn’t), to prices being discontinuous around the cap. As demand shifts from left to right, prices behave normally and rise as they normally would, until abruptly jumping higher once the capped producer is removed. In any case, price is more volatile than it would otherwise be…but, and this is important, it is never lower in a market where some or all of the suppliers are capped, than it is in an uncapped market. At best, prices are the same if the caps aren’t in play. At worst, a combination of shortages and higher prices obtain.
Speaking of shortages…it seems that people are growing calmer about the chances of a bad energy outcome over the winter in Europe. This seems, to me, to be related to the fact that inventories of gas are reasonably flush thanks to conservation efforts and vigorous efforts to replace lost Russian pipeline supply (see Chart, source Gas Infrastructure Europe via Bloomberg).
That’s great, but the problem is that since the pipelines are not flowing Europe needs more gas going into the winter than they otherwise would have – because it’s not being replenished by pipeline during the winter, either. We certainly hope that Europe doesn’t run out of heat this winter, but the level of gas inventories is not exciting.
Putting downward pressure on both of these markets, but especially Crude, is the idea that the world will enter a global recession in 2023. As I’ve been saying since early this year, that’s virtually a sure thing: we’ve never seen interest rates and energy prices rise this much and not had a recession. But I have thought that the recession would be relatively mild, a ‘garden variety’ recession compared to the last three we’ve had (the tech bubble implosion, the global financial crisis, and the COVID recession). What worries me a bit is that the consensus is now moving to that conclusion. It seems that most forecasts are for a mild recession (although predictably, economists are all over the map on inflation depending on the degree to which they understand that inflation is a monetary phenomenon and not a growth phenomenon). I’m still in that camp, but that concerns me, because the consensus is usually wrong.
[1] In round 2, after oil delivery from Russia is switched to the uncapped markets, the available price in the capped market will need to be appreciably above the market clearing price in the uncapped market in order to cause the switch back.
The Coming Rise in Money Velocity
As M2 money growth soared throughout the COVID and post-COVID period of direct stimulus check-writing funded by massive quantitative easing (QE), monetarism novices thought that this would not result in inflation because money velocity simultaneously collapsed. Consequently, they argued, M*V was not growing at an outrageous rate.
There was precedence for such optimism. In the Global Financial Crisis of 2008-09, money supply grew rapidly with the onset of QE and money velocity declined, never to recover. The chart below shows in a normalized fashion the rise in M, the decline in V, and the relative quiescence of MV/Q, which is of course P by definition as long as you choose your Ms, Vs, and Qs right.
A similar thing happened in this episode, so why would this be any different?
There are many reasons why these episodes are different. To name a few:
- The absolute scale of the rise in M2 was 2.5x the rise in 2007-2010, and that’s being generous since that measures the growth in 2007-2010 starting almost 2 years before the first QE in November 2008 compared to only 15 months in the second case.
- As I’ve written previously, QE in the first case was directed at banks; at the same time that the Federal Reserve was adding reserves it was also paying banks interest on reserves – because the point was to strengthen banks, not consumers.
- 5y interest rates came into 2008 at 3.44%; they came into 2020 at 1.69%. Since velocity is most highly correlated to interest rates, there was less room for this factor to be a lasting downward influence on velocity (after the crisis began in 2008, 5y Treasury rates never exceeded 3% again except for a few days in 2018).
- Bank credit growth never stopped in the 2020 crisis, while it contracted at a 5% rate in the 2008 crisis (see chart, source Board of Governors of the Fed).
The monetarist novices (you can tell they’re novices because they say things like “Friedman said velocity was constant,” which is false, or “velocity is just a plug number [true] and has no independent meaning of its own [false]”) insisted that velocity was in a permanently declining state and that there was no reason at all to expect it to ever “bounce.” After all, it bounced only slightly after the GFC; why should it do so now?
But after 2008, as I noted, interest rates bounced only briefly before declining again…with the added phenomenon that some global debt came to bear negative yields, calling into fair question whether there was in fact any natural “bottom” to velocity if interest rates are the main driver! And velocity, obediently, dripped lower as well.
There is at least one other big driver to money velocity, although it is rarely important and almost never for very long. And that is economic uncertainty, which creates a demand to carry excess cash balances (implying lower money velocity). A model driven (mainly) by rates and a measure of uncertainty has done a pretty good job at explaining velocity over time (see chart, source Enduring Investments), including explaining the collapse in velocity during the COVID crisis out-of-sample.
Now, explaining velocity is a helluva lot easier than predicting it, because it isn’t easy to predict interest rates. Nor is it easy to predict the precautionary demand for money – but at least we can count on that being somewhat mean-reverting. The latest point from the model shown above uses current data, and suggests (largely because of the rise in interest rates, but also because precautionary balances are declining) that money velocity should bounce. Not that the model predicts it will happen this week, but it should not be surprising when it does.
A rise in velocity would be a really bad thing, because the money supply is very unlikely to decline very far especially while bank credit growth continues to grow. The only reason we have been able to sustain 6% or 8% money growth for a very long time has been because we could count on velocity to keep declining with interest rates. If money growth ticks up at, say, a mere 6% while money velocity rises 5%, then nominal GDP is going to rise 11%…and most of that will be in prices.
Now, this is a very slow-moving story. I mention it now for one specific reason, and that is that we are almost certain to see a rise in velocity in Q2 when the GDP figures come out in late July. That’s because money growth for the quarter has been very slow so far. So far, the Q2 average M2 is 0.06% higher than the Q1 average. My best wild guess is that we will end up with an 0.5% annualized q/q growth rate. The Atlanta Fed GDPNow model estimates 0.25% GDP growth in Q2 (the Blue Chip Consensus is still at 3%). And if the inflation market is right, Q/Q inflation in Q2 will be about 11.7%. That’s CPI, so let’s be generous and say 9%. We don’t know all of these numbers, but we know 2/3 of all of them. Let’s use the Blue Chip consensus for GDP and assume M2 doesn’t spike next month and the price level doesn’t collapse. Then:
If that happened, the increase would be the largest quarterly jump in money velocity – absent the reactionary bounce in 2020Q3 after the 20% plunge in 2020Q2 – since 1981. And here’s the rub: because of the mathematics of declines and recoveries, that would still leave us with velocity that prior to 2020 would have been an all-time record low.
Does this matter? Not if you believe the monetarism dabblers, who will say this is a mechanical adjustment that will soon be reversed as velocity continues its long slide to oblivion. Nor will it matter to the Fed, who at best will take executive notice of the fact before ignoring it since they aren’t monetarists any longer. But for those who think that inflation comes from too much money chasing too few goods? It’s scary.
Why Roughly 2.25% is an Equilibrium Real Rate
Recently, Fed officials have taken to discussing “long-term equilibrium” interest rates as a way of indicating to the market where interest rates might ultimately be heading. It is not exactly a terrifying prospect. The Fed seems to collectively believe that the “neutral” short-term nominal interest rate is in the 2.50%-2.75% range; some fear that the Fed funds target right may have to be lifted “modestly” above this level for a time. This seems hard to believe, with inflation running with an 8% handle – such an overnight rate would equate to an annual 5-6% incineration of purchasing power. The only way this could be considered “neutral” is if one begs the question by asserting contrary to evidence that the long-run equilibrium inflation rate is around 2%-2.25%.
I have noted repeatedly over the last year or so why it is unlikely in my opinion that the current equilibrium for inflation is in the 2% range; I feel it is closer to 4%-5% in the medium-term. But if an observer has a model which has been ‘trained’ on data from the last thirty years, the model will assuredly tell you that any time inflation deviates from 2%, it comes back to 2%. In fact, any model which did not produce that prediction would not have been considered a good model: it would have made predictions which, for 30 years, would have been noticeably incorrect from time to time. Ergo, all surviving models will view something like 2% as an attracting level for inflation, and we know the Fed continues to believe this. So, evidently, do many other economists. I keep showing this following chart because I think it’s delicious. Take today’s level; take the level your model says is a self-enforcing equilibrium, and draw a straight line. That’s your forecast. You too can be a million-dollar Wall Street economist.
Faced with awful predictions from this cadre of models, one solution is to consider why they had bad predictions, and attempt to develop models that would perform on data from the 1970s and 1980s as well. A more attractive solution, from an institutional perspective, is to blame model-exogenous events. That is, “the model is fine; who could have foreseen that supply chain issues would have triggered such a large inflation?” And so, we preserve the FOMCs ability to continue making terrible forecasts.
Similarly, Minneapolis Fed President Neel Kashkari stated not too long ago that the Fed may have to “push long-term real rates into restrictive territory.”[1] This continues the Fed’s error of obsessing on the price of liquidity rather than its quantity, but that isn’t the point I am making here. Kashkari made a different error, in an essay posted on the Minneapolis Fed website on May 6th.[2] He claimed that the neutral long-term real interest rate is around 0.25%, which conveniently is where long-term real rates are now.
However, we can demonstrate that logic, reinforced by history, indicates that long-term real rates ought to be in the neighborhood of the economy’s long-term real growth rate potential.
I will use the classic economist’s expedient of a desert-island economy. Consider such an island, which has two coconut-milk producers and for mathematical convenience no inflation, so that real and nominal quantities are the same. These producers are able to expand production and profits by about 2% per year by deploying new machinery to extract the milk from the coconuts. Now, let’s suppose that one of the producers offers to sell his company to the other, and to finance the purchase by lending money at 5%. The proposal will fall on deaf ears, since paying 5% to expand production and profits by 2% makes no sense. At that interest rate, either producer would rather be a banker. Conversely, suppose one producer offers to sell his company to the other and to finance the purchase at a 0% rate of interest – the buyer can pay off the loan over time with no interest charged. Now the buyer will jump at the chance, because he can pay off the loan with the increased production and keep more money in the bargain. The leverage granted him by this loan is very attractive. In this circumstance, the only way the deal is struck is if the lender is not good at math. Clearly, the lender could increase his wealth by 2% per year by producing coconut milk, but is choosing instead to maintain his current level of wealth. Perhaps he likes playing golf more than cracking coconuts.
In this economy, a lender cannot charge more than the natural growth in production since a borrower will not intentionally reduce his real wealth by borrowing to buy an asset that returns less than the loan costs. And a lender will not intentionally reduce his real wealth by lending at a rate lower than he could expand his wealth by producing. Thus, the natural real rate of interest will tend to be in equilibrium at the natural real rate of economic growth. Lower real interest rates will induce leveraging of productive activities; higher real interest rates will result in deleveraging.
This isn’t only true of the coconut economy, although I would strongly caution that this isn’t exactly a trading model and only a natural tendency with a long history. The chart below shows (1) a naïve real 10-year yield created by taking the 10-year nominal Treasury yield and subtracting trailing 1-year inflation, in purple; (2) a real yield series derived from a research paper by Shanken & Kothari, in red; (3) the Enduring Investments real yield series, in green, and (4) 10y TIPS, in black.
The long-term averages for these four series are as follows:
- Naïve real: 2.34%
- Shanken/Kothari: 3.13%
- Enduring Investments: 2.34%
- 10y TIPS: 1.39%
- Shanken/Kothari thru 2007; 10y TIPS from 2007-present: 2.50%
It isn’t just a coincidence that calculating a long-term average of long-term real interest rates, no matter how you do it, ends up being about 2.3%-2.5%. That is also close to the long-term real growth rate of the economy. Using Commerce Department data, the compounded annual US growth rate from 1954-2021 was 2.95%.
It is generally conceded that the economy’s sustainable growth rate has fallen over the last 50 years, although some people place great stock (no pun intended) on the productivity enhancements which power the fantasies of tech sector investors. I believe that something like 2.25%-2.50% is the long-term growth rate that the US economy can sustain, although global demographic trends may be dampening that further. Which in turn implies that something like 2.00%-2.25% is where long-term real interest rates should be, in equilibrium.[3] Kashkari says “We do know that neutral rates have been falling in advanced economies around the world due to factors outside the influence of monetary policy, such as demographics, technology developments and trade.” Except that we don’t know anything of the sort, since there is a strong argument against each of these totems. Abbreviating, those counterarguments are (a) aging demographics is a supply shock which should decrease output and raise prices with the singular counterargument of Japan also happening to be the country with the lowest growth rate in money in the last three decades; (b) productivity has been improving since the Middle Ages, and there is no evidence that it is improving noticeably faster today – and if it did, that would raise the expected real growth rate and the demand for money; and (c) while trade certainly was a following wind for the last quarter century, every indication is that it is going to be the opposite sign for the next decade. It is time to retire these shibboleths. Real interest rates have been kept artificially too low for far too long, inducing excessive financial leverage. They will eventually return to equilibrium…but it will be a long and painful process.
[1] https://www.reuters.com/business/finance/feds-kashkari-we-may-have-push-long-term-real-rates-into-restrictive-territory-2022-05-06/
[2] https://www.minneapolisfed.org/article/2022/policy-has-tightened-a-lot-is-it-enough
[3] The reason that real interest rates will be slightly lower than real growth rates is that real interest rates are typically computed using the Consumer Price Index, which is generally slightly higher than the GDP Deflator.
Shortages are Unmeasured Inflation
Recently, I’ve been saying occasionally that “shortages are unmeasured inflation.” In some conversations I have had, it became apparent to me that people were taking this statement as being a throwaway line: “inflation is bad, shortages are bad, therefore they’re kinda the same.” But what I mean is actually more profound than that, and so I figured I would explain and illustrate, and hopefully thereby to convince.
Let’s use some charts.
What has happened since the large increase in federal spending and transfer payments were implemented in several waves since the shutdown began is that demand in many product markets has shifted outward. This implies that output “Q” moves from c to d while the price level “P” moves from a to b. [1]
So a strong increase in demand causes an increase in the quantity exchanged at the new equilibrium, and an increase in the price of the good or service at that equilibrium. This is the nice, smooth, continuous markets, instantaneous-adjustment picture from Econ 101. It’s also not the way the real world works, especially with large shifts in demand.
If price only adjusts partially, maybe if “anchored inflation expectations” or a fear of being accused of gouging restrained vendors from raising prices enough to ration the available supply, then a shortage results. This is the same thing which occurs classically if a price cap is instituted from the outside. Now price moves up from a only to b’, but the quantity demanded at that price is at d’. Thus, the bracket on the chart below shows the size of the shortage at that price, where consumers want d’ but suppliers can’t/won’t provide that much.
Note that this shortage is a direct substitute for the increase in price that would otherwise happen if prices could instantly and fully adjust. Moreover, the picture is somewhat worse in the short-term because the supply curve – in the short-term – is much more inelastic at some point (because, for example, no matter how high the price gets we can’t deliver more used cars in the short run). So, in the picture below the short-run supply curve in blue implies that the large increase in demand pushes prices to b’ with output only at d’, until supply eventually adjusts to the long-run supply curve S(lr), when we end up in the new market-clearing equilibrium.
In this case, the difference between b and b’ is “transitory” inflation, caused by temporary supply constraints. But note that in this picture, there is no perceived shortage. The market clears at b’ and d’. In other words, the conditions leading to the “transitory” increase inflation are not the same ones leading to the shortages.
We can combine these; if in the last picture above vendors also constrain prices to b, then there is a shortage as the quantity demanded stays up at d rather than at the market-clearing level d’. But, again, in that instance the shortage reflects the fact that prices should have adjusted to b’ but did not. Also in that case, it would be inaccurate to claim that the inflation was transitory, since prices should remain at b even when the supply eventually adjusts to the long-run equilibrium. It would be the shortage that was transitory.
In theory, if we knew the shapes of the curves of supply and demand for each product market, we could estimate how much higher prices would be at equilibrium and therefore how much additional inflation the shortage implies. We could directly translate the shortages to an “equilibrium” price level and therefore inflation. It strikes me as plausible that we could develop a rough estimate of such a number, but I leave that to the PhD candidates looking for dissertation topics. In the meantime, just remember that with inflation over 5% presently and shortly headed above 6%…the inflation rate is understated, and we know that because there are lots of shortages.
[1] If the deficits, funded by Fed purchases of Treasuries, had just offset the loss in incomes due to the shutdown – perfectly, across all individual markets – then there would have been no demand shift and no net change in output or prices. And if the deficits had not been accompanied by an increase in the Fed balance sheet, then it would have been individuals buying the bonds and so the only effect would have been because the marginal propensity to consume of the people receiving transfer payments is higher than the marginal propensity to consumer of the people buying the bond issuance. But in this column I’m trying not to muddy the discussion with the argument of whether we need both fiscal and monetary stimulus to cause inflation. I’m just focused on the narrow question of what it means when I say “shortages are unmeasured inflation.”
Why We’re Wrong About Restaurants
Figuring out the macro impact of the virus, while not easy, is in some ways easier than figuring out a lot of the micro. In some cases the impact seems pretty obvious, and probably is: airlines are likely to carry fewer passengers, and more of them will be business travelers, for a while (resulting, by the way, in higher airfares in CPI). But some of the effects are much harder to figure out than we think, and a lot of it comes down to the fact that people who are idly speculating about these things tend to be pretty poor about defining what the substitutes are for any product or service.
Actually, the question of ‘what is a substitute’ turns out to be hugely important in economic modeling, because it directly impacts the question of demand elasticity. If I am the only person who sells widgets, and you need a widget, then I probably have a lot of control over what you pay. But if someone else sells something that works about as well as a widget (but isn’t a Widget™), then I as the supplier likely have a lot less flexibility and I face a more elastic demand curve. This is one reason that salespeople are taught to remember that the customer doesn’t want a quarter-inch drill bit; they want a quarter-inch hole. In a more formal setting: it is enormously important in antitrust economics that the market is defined clearly when considering if a firm is monopolizing or attempting to monopolize[1], so much so that there is an index called the Herfindahl-Hirschman Index with which industry concentration can be expressed. But I digress.
We read that many restaurants will fail as a result of the COVID-19 crisis, because quite aside from the question of the financial damage done to the restaurant owner from a two-month hiatus in revenues there is the question of “will people even come back?” And, if people do come back, but the restaurant-owner can only fit half as many people in the restaurant due to social distancing, then many restaurants can’t survive. Right?
So we are told, but there are a ton of assumptions there and some of them don’t hold. One of the biggest assumption is the question of what consumers use as a substitute for restaurant meals. With airlines, there is a clear substitute for the vacation traveler and that’s the automobile. Moreover, a vacation is not a necessity per se. But everyone needs to eat, so we can say with some confidence that if the average American ate 2.5 meals per day before the crisis they will probably eat 2.5 meals per day after the crisis. Somehow, they need to get those meals. If they are not going to restaurants for some of those meals, what are the alternatives? The argument that restaurants will fail hinges partly on the idea that these alternatives are convenient enough and enough competition for restaurant meals that consumers will eschew eating out and so restaurants won’t be able to sell their product. But will they? The alternatives to a restaurant meal are (a) a meal cooked at home or (b) a meal delivered. Many restaurants might fail for financial reasons, but that happens all the time in the food preparation biz. The question is whether the total number of restaurants in the country will be dramatically lower in the post-virus world. If so, it means that people are choosing en masse to make a significantly higher percentage of their meals at home. Anyone reading this who is doing a lot of their own cooking these days will realize why that’s probably not a tenable outcome as long as we continue to need two incomes in most families! Some, surely, will cook more. But when this is over, I suspect that meals not prepared at home will be a similar portion of our diets as it was before.
“Meals not prepared at home” includes both restaurant meals and delivery meals. Since the total number of meals consumed will be roughly the same, I think we’ll see a bit more home-cookin’ and a lot more delivery. It will be the restaurants, even those that did not previously deliver, cooking those meals. Maybe more delivery-only restaurants will start up. But I really think that we will see almost as many restaurants a year from now as we do now.
A separate question is what happens to the price of a meal-not-cooked-at-home, and it seems to me that the answer must be that it goes a lot higher. A delivered meal requires more manpower (for delivery), especially if delivery is going to be efficient at all. And in-restaurant meals (for those dinners, like your anniversary dinner, for which there are no good substitutes) are going to be higher-priced both because the demand curve will be more inelastic in the same way that the demand curve for business air passengers is more inelastic, and because the supply will be constrained. But my point is that if the restaurant used to plate 100 meals per hour, they’ll still plate pretty close to 100 meals per hour. It’s just that 50 of those meals will be going out the door.
Is there a substitute for movie theaters? Absolutely, and it was already winning. Good-bye movie theaters (although I have seen something about drive-ins making a comeback). A substitute for sports venues? Not so much, so I think we’ll see some innovation about how we safely attend such events but we haven’t seen the last of major league baseball at Citi Field or rugby at Twickenham. I think that international visitors to Disney World will probably decline, but domestic visitors will probably increase, as Disney for the latter is a substitute for an island vacation. But those islands that depend on tourism – there will be some pain there as there aren’t many convenient ways to get to Martinique that don’t involve flying.
But while I’m sure some restaurants will close because they cannot figure out delivery or because their product doesn’t translate well to delivery (see this story about a high-end restaurant that is facing this dilemma), I think consumption of meals-not-cooked-at-home will ensure that we will have a similar number of restaurants in the future. The broader point is this: be careful when you’re thinking about the damage that certain businesses will experience. Be sure to think about what the market for the good or service is, and what the relevant competitors are. Again, this doesn’t mean that existing companies will always survive, but if you know that the market for (for example) automobiles is still going to be there then there will be companies that serve that market. If they are different companies than today’s companies, that’s just creative destruction and it isn’t a bad thing for the consumer. (And, personal pitch: if you or your company needs help navigating these waters, visit our new website at https://www.enduringinvestments.com and drop me a line.)
[1] See Tasty Baking Company and Tastykake, Inc. v. Ralston Purina, Inc. and Continental Banking Co. (1987) in which the plaintiffs argued that the relevant market was premium snack cakes and pies and defendants argued that their products competed in the market for ‘all sweet snacks,’ because obviously their combination was less dominant if there were lots of substitutes.
COVID-19 in China is a Supply Shock to the World
The reaction of much of the financial media to the virtual shutdown of large swaths of Chinese production has been interesting. The initial reaction, not terribly surprising, was to shrug and say that the COVID-19 virus epidemic would probably not amount to much in the big scheme of things, and therefore no threat to economic growth (or, Heaven forbid, the markets. The mere suggestion that stocks might decline positively gives me the vapors!) Then this chart made the rounds on Friday…
…and suddenly, it seemed that maybe there was something worth being concerned about. Equity markets had a serious slump yesterday, but I’m not here to talk about whether this means it is time to buy TSLA (after all, isn’t it always time to buy Tesla? Or so they say), but to talk about the other common belief and that is that having China shuttered for the better part of a quarter is deflationary. My tweet on the subject was, surprisingly, one of my most-engaging posts in a very long time.
It has been so long since we have seen a supply shock that we have forgotten what they look like. China Inc shutting down is a supply shock. Supply shocks are inflationary, not disinflationary. Carry on.
— Michael Ashton (@inflation_guy) February 24, 2020
The reason this distinction between “supply shock” and “demand shock” is important is that the effects on prices are very different. The first stylistic depiction below shows a demand shock; the second shows a supply shock. In the first case, demand moves from D to D’ against a stable supply curve S; in the latter case, supply moves from S to S’ against a stable demand curve D.
Note that in both cases, the quantity demanded (Q axis) declines from c to d. Both (negative) demand and supply shocks are negative for growth. However, in the case of a negative demand shock, prices fall from a to b; in the case of a negative supply shock prices rise from a to b.
Of course, in this case there are both demand and supply shocks going on. China is, after all, a huge consumption engine (although a fraction of US consumption). So the growth picture is unambiguous: Chinese growth is going to be seriously impacted by the virtual shutdown of Wuhan and the surrounding province, as well as some ports and lots of other ancillary things that outsiders are not privy to. But what about the price picture? The demand shock is pushing prices down, and the supply shock is pushing them up. Which matters more?
The answer is not so neat and clean, but it is neater and cleaner than you think. Is China’s importance to the global economy more because of its consumption, as a destination for goods and services? Or is it more because of its production, as a source of goods and services? Well, in 2018 (source: Worldbank.org) China’s exports amounted to about $2.5trillion in USD, versus imports of $2.1trillion. So, as a first cut – if China completely vanished from global trade, it would amount to a net $400bln in lost supply. It is a supply shock.
When you look deeper, there is of course more complexity. Of China’s imports, about $239bln is petroleum. So if China vanished from global trade, it would be a demand shock in petroleum of $240bln (about 13mbpd, so huge), but a bigger supply shock on everything else, of $639bln. Again, it is a supply shock, at least ex-energy.
And even deeper, the picture is really interesting and really clear. From the same Worldbank source:
China is a huge net importer of raw goods (a large part of that is energy), roughly flat on intermediate goods, and a huge net exporter of consumer and capital goods. China Inc is an apt name – as a country, she takes in raw goods, processes them, and sells them. So, if China were to suddenly vanish, we would expect to see a major demand shock in raw materials and a major supply shock in finished goods.
The effects naturally vary with the specific product. Some places we might expect to see significant price pressures are in pharmaceuticals, for example, where China is a critical source of active pharmaceutical ingredients and many drugs including about 80% of the US consumption of antibiotics. On the other hand, energy prices are under downward price pressure as are many industrial materials. Since these prices are most immediately visible (they are commodities, after all), it is natural for the knee-jerk reaction of investors to be “this is a demand shock.” Plus, as I said in the tweet, it has been a long time since we have seen a serious supply shock. But after the demand shock in raw goods (and possibly showing in PPI?), do not be surprised to see an impact on the prices of consumer goods especially if China remains shuttered for a long time. Interestingly, the inflation markets are semi-efficiently pricing this. The chart below is the 1-year inflation swap rate, after stripping out the energy effect (source: Enduring Investments). Overall it is too low – core inflation is already well above this level and likely to remain so – but the recent move has been to higher implied core inflation, not lower.
Now, if COVID-19 blossoms into a true global contagion that collapses demand in developed countries – especially in the US – then the answer is different and much more along the lines of a demand shock. But I also think that, even if this global health threat retreats, real damage has been done to the status of China as the world’s supplier. Although it is less sexy, less scary, and slower, de-globalization of trade (for example, the US repatriating pharmaceuticals production to the US, or other manufacturers pulling back supply chains to produce more in the NAFTA bloc) is also a supply shock.
A Generous Fed Isn’t Really the Good News it Sounds Like
I understand why people are delighted about Powell’s remarks yesterday, about how the Fed would need to see a significant and sustained increase in inflation before hiking rates again. This generation, and the last, does not see inflation as a significant threat, nor a significant cost should it get going, and believes firmly that the Fed can easily squelch it if it gets going. (They believe this because, after all, the Fed told them so).
Older investors might be more reticent to believe that there’s a pony in there somewhere, since the evidence suggests that not only does inflation erode purchasing power (thereby demanding even more nominal return be provided by portfolios that are already overstretched valuation-wise) but it also ruins the diversification effect of bonds relative to stocks. The main reason that 60:40 is a dramatically lower risk portfolio (and more efficient in an investing sense) than 100% stocks is that stock and bond returns have tended to be inversely correlated for a long time. When stocks go up, bonds go down, in general (and vice-versa). But that’s because they have inverse sensitivities to the economic growth factor. In recent years, that has been the only factor that matters, but stocks and bonds have the same sensitivity to the inflation factor: when inflation goes up, both stocks and bonds tend to decline (and vice-versa). Consequently, when inflation becomes an important element in investors’ calculations the correlation of stocks and bonds tends to be positive and in the immortal words of Billy Joel in “Goodnight Saigon,” “We would all go down together.” Along these lines I recently prepared this chart for Real Asset Strategies,[1] illustrating that when inflation is over about 2.5%, correlations tend to flip. This is a 3-year average of y/y inflation (and shown on the chart as inflation minus 2.5% so the zero line is what matters, not the line at 2.5%) versus 3-year correlations; the point is that you don’t need 4% inflation to drastically change the value of the 60:40 portfolio.
I also think that people give the Fed much more credit for their ability to squelch inflation – which after all they haven’t had to do for more than 30 years after spending 15 years squelching the last round – than they deserve. But that’s a ‘show me’ situation and it’s hard to prove my suspicion that they won’t be so successful when push comes to shove.
So, I understand why people are partying about a Fed that is even looser than it had been. I don’t think that’s the correct response, but I understand it.
I also understand why people are somewhat morose about trade frictions. It isn’t for the right reason, that in the long run it will hurt real growth a smidge and increase inflation a smidge-and-a-half, but because they think it will have a drastic effect on near-term growth. That’s why everyone gets so excited about any inkling the US and China are nearing a trade détente and so depressed when it looks like they aren’t. We are told that the current global slowdown is being caused largely by the trade war.
In my view that’s nonsense. The global economy has been expanding for a decade on exceptionally loose liquidity but no tree grows to the sky. The global economy was slowing well before the trade frictions could possibly have had any impact. But it is hard to convince people of that, because everyone knows that:
GDP = C + I + G + (X-M),
or consumption plus investment plus government spending plus trade. And we learned in school about Ricardian comparative advantage and how trade enriches (or anyway, can enrich) both parties at the same time. So if China doesn’t import anything from the US and doesn’t export anything to the US, growth is going to be crushed, right?
But that’s not how trade works. Frankly, that’s not how anything in the GDP equation works. If you remove the final term, you don’t reduce GDP by (X-M). Sure, if this was an algebra problem you would, but it’s not. In the real world, what you lose from trade gets partially replaced by an increase in consumption, investment, or government. Just as I pointed out last year with soybeans, if China buys zero from us it means they have to buy them from someone else, which means that supplier doesn’t have them to sell to one of their traditional customers…who then buys them from us. Incidentally, neither beans nor corn went to zero after mid-2018 (see chart, source Bloomberg, normalized to December 2017=100).
The rest of trade works the same way if the two parties are “internal customers” and “external customers.” Though there will always be winners and losers, if we don’t have international trade then we won’t have a destination for our merchandise overseas…but we will also have consumers who don’t have Chinese goods to buy and so need to buy something from a domestic producer instead. This is not a zero sum game; it clearly results in a loss for all players. But the order of magnitude of this loss in the short run is not very big at all, especially for a country with a large fraction of its domestic production going to domestic consumption, as in the US but not even for the world at large. The world economy has lots of reasons to slow and go into recession, and trade frictions are one of those reasons, but certainly not the only one and not even the largest reason.
An overreaction by markets to anything in a stream of economic news is not unique or new, of course; those overreactions won Robert Shiller a Nobel Prize after all for his work pointing out the “excess volatility puzzle” as an early highlight of the nascent field of behavioral economics. But there’s a good reason to ignore most of these wiggles and focus on the long-term effect of these developments. Which, in the case of both the general climate of trade and the Fed’s reaction function to inflation, are negatives for both stocks and bonds.
[1] As part of Enduring Intellectual Properties’ investment in Real Asset Strategies, I serve as Director of Research for the firm. Real Asset Strategies LLC offers liquid real asset strategies focused on diversification benefits and inflation protection at reasonable fees.