E-piphany

Below is a summary of my post-CPI tweets. You can (and should!) follow me @inflation_guy, but to get these tweets in real time on CPI morning you need to subscribe to @InflGuyPlus by going to the shop at https://inflationguy.blog/shop/ , where you can also subscribe to the Enduring Investments Quarterly Inflation Outlook. Sign up for email updates to my occasional articles here. Individual and institutional investors, issuers and risk managers with interests in this area be sure to stop by Enduring Investments! Check out the Inflation Guy podcast!

• It’s #CPI Day – and this one finishes up the book for 2022.
• I am doing the walk-up differently today. I’m doing it as a thread on the night before, which I’ll re-tweet in the morning. I’m usually doing the analysis in the evening…why wait?
• Today’s number, or I guess really we can say starting with October or November, starts the interesting part of the inflation cycle.
• When inflation was going up, excuses abounded but the real debate was WHEN the peak was going to be, and HOW HIGH only to a lesser extent. Now that inflation appears to be clearly decelerating, the much more important debate is: where is it decelerating to?
• If inflation drops back to 2%, and becomes inert at that level again, then the Fed will deserve considerable laurels. If inflation instead drops to 4% and appears resistant to a drop below that, then a much more interesting debate will ensue.
• I think it should be clear that I am in the latter camp.
• The other interesting thing that we’re going to see, and are already seeing, is manifestation of the basic tricks of the trade of macro economists.
• Trick 1 is to assume that everything returns to the mean. Most things do, eventually, return to the mean – so if you are wrong on the timing, you’ll probably eventually be right. Economists love to forecast returns to the mean.
• Economists though are very bad at forecasting departures AWAY from the mean, which is why there were so many forecasts of “transitory” this cycle.
• Since they didn’t see it coming, it must have been a random perturbation (because that’s how their models work). But it’ll all go back to the mean and all is right with the world. Or so goes the assumption.
• Trick 2 is to assume that the mean doesn’t change, or changes pretty slowly. In econometrics terms, the distribution is ‘stationary.’ If you’re going to forecasts returns to the mean, it is fairly important that the ‘mean’ is known or knowable and doesn’t move a lot.
• The problem in inflation is that the (unobservable) mean of the distribution never appeared to be very stable until the mid-1990s; the hypothesis is that this anchoring happened because of “anchored inflation expectations.”
• (A member of the Fed’s own research staff tore apart that notion in a devastating article a couple of years ago, but the Fed promptly ignored him because if he was right it’s really bad for forecasting the way that they like to forecast: everything returns to the mean.)
• Getting to Thursday’s CPI figure, we can see these tricks in play in the economist forecasts.
• As an example, one of the forecasts I saw from a large bank had drags calculated from Used Cars (and New Cars), a deceleration in shelter costs, a drag from airfares due to lower jet fuel costs, and a drag from health insurance. But what about accelerations?
• Do you really think that NOTHING will accelerate, or are all of those pre-defined as “one-offs”?
• It reminds me a little of what Rob Arnott says about the S&P earnings “ex-items”: any one company it might make sense to ex- the unusual events. But in aggregate, some level of unusual events is usual. So it is with inflation.
• There will be some ups. So my forecasts are a little higher than others’, because I anticipate there will be some surprises.

• Where would those surprises come from? Wage growth is strong, and that pushes up on prices in hospitality, domestic manufacturing, food away from home, and even shelter.

• I also don’t think that airfares will be the drag that’s implied by jet fuel. Here’s the regression that would make you think they WOULD.

• But here’s the one that makes you think maybe not. Airlines tend to push prices higher when there are spikes in jet fuel costs, but they don’t necessarily lower them very fast when jet fuel prices decline. And did I mention wage pressures? Airlines feel them.

• I do think that used car prices will drag again, although the CPI has been falling a little faster than the Black Book and Mannheim indices would suggest they should. But I don’t see a strong argument for New Car prices to decline.
• New Cars are in black in this chart, while Used are in blue. New car prices are up 20%, while used are up 40%, since the end of 2019. And the money supply is up around 40%. That doesn’t mean new car prices won’t decline, but it doesn’t look like a slam dunk to me.

• Finally, a point I’ve been making recently on a longer-term horizon viewpoint. Markets are fully priced for inflation to totally and almost immediately mean-revert. Large declines in breakevens, especially short BEI. Some of that is the gasoline slide. Not all of it.

• The short end of the inflation swap curve has NSA inflation at -0.38% m/m in December, +0.37% in Jan, +0.33% in Feb, and 0.30% in March. And that’s the last 0.3% print we see. According to inflation swaps, y/y inflation will be at 2% in June.
• Even if I am wrong about inflation staying around 4-5%, you have a 2% cushion to bet that way. (I think I used an unfortunate analogy a few days ago saying that if you give me 21 points I’ll take TCU over Georgia, but you get my point.)
• Ergo, for choice I’d be long breakevens going into this number.
• The response in the stock market will be interesting. If the number is as-expected or better, I would think stocks will try and scream higher on the theory that the Fed can back off. The problem is that folks are already long for that, I sense.
• So I’d probably sell that pop, especially because earnings may be a hurdle in the near future, though you have to be cognizant of the 200-day moving average in the S&P. The mo-mo crowd will try to get some prints above that so I’d be cautious.
• What about on a strong CPI? Few seem to be thinking/talking of that, which means to me that folks are a little naked there. Do I think it would change the Fed trajectory? Not from what the Fed is SAYING they’re doing, but from what the market is pricing – yes.
• As I said, this is the interesting part of the inflation cycle. Buckle up.
• At 8:30ET, I’ll be pulling the data in & will post charts and #s – then retweet some of those charts w/ comments plus other charts. Around 9:30ish, I will have a private conference call for subscribers where I’ll quickly summarize the numbers.
• Pre-release, both stocks and bonds are loving this number! May be that some are reading into the fact Biden has a speech this morn including inflation as a topic, and perhaps he wouldn’t if the number was bad. But even if it is, he can focus on y/y so not sure that means much…
• That’s all for now. Good luck!

---

• m/m CPI: -0.0794% m/m Core CPI: 0.303%
• Last 12 core CPI figures

• Overall, highest core number in 3 months, but clearly in a down trend. I think lots of people would be DELIGHTED with 3.6% annualized compared with where we have been, but that’s closer to what I am expecting than what the market/Fed is looking for.
• M/M, Y/Y, and prior Y/Y for 8 major subgroups

• Interesting thing is apparel, up for the second month in a row. Apparel is an almost pure import, so if it’s up then either (a) the recent dollar weakness is already affecting prices or more likely (b) there is pricing power at retail, and the markdowns for Christmas were lower.
• Core Goods: 2.15% y/y Core Services: 7.05% y/y

• The story continues to be bifurcated and we will look further at the four-pieces. More important than the fact that services are trending and goods are deflating, is whether the services part was all rents.
• Here is my early and automated guess at Median CPI for this month: 0.378%

• Clearly good news! Lowest median m/m in quite some time. So core was higher, but median lower. THIS is positive. And as I said, this is the interesting part now: inflation is decelerating, but why and how fast and how far? Median clearly shows it is.
• Primary Rents: 8.35% y/y OER: 7.53% y/y

• Further: Primary Rents 0.79% M/M, 8.35% Y/Y (7.91% last) OER 0.78% M/M, 7.53% Y/Y (7.13% last) Lodging Away From Home 1.5% M/M, 3.2% Y/Y (3.2% last)
• Although the rent data is clearly bad news, there has been a strong campaign against this data to weaken its importance by claiming it’s just really lagged. That’s partly true but the recent research on the subject has enormous error bars for short-term forecasts so…
• Some ‘COVID’ Categories: Airfares -3.12% M/M (-3.02% Last) *** Lodging Away from Home 1.47% M/M (-0.71% Last) *** Used Cars/Trucks -2.55% M/M (-2.95% Last) *** New Cars/Trucks -0.06% M/M (0.04% Last)
• So, I was ‘on’ core even though I was wrong on airfares (it was weak, despite the fact that every fare I saw in December was about 2x normal). Used cars was the predicted drag, and New cars was not…but I was low on rents. That’s the ‘away from mean surprise’.
• Incidentally, Lodging Away from Home was quite strong – and is one of those core-services-ex-rents that is driven a lot by wages.
• Piece 1: Food & Energy: 9.31% y/y

• Piece 2: Core Commodities: 2.15% y/y

• Piece 3: Core Services less Rent of Shelter: 6.34% y/y

• …and here is the spoiler: it wasn’t all rents. Core services less rents still strong. I’ll drill down further in a bit.
• Piece 4: Rent of Shelter: 7.59% y/y

• So, the swap market gets closest-to-the-pin on headline (SA). -0.079% was the figure, a bit lower than consensus econs and a fair bit lower than me. On Core, econs and I were both pretty close as it was right around 0.3% (0.303%).
• I had managed to talk myself into the idea that food and energy would be a bit less of a drag than my model said, but food wasn’t up as much as it has recently been. Ergo, right on core and off on headline.
• Interesting story in Medical Care, which has been a drag recently because of the huge adjustment to insurance company margins (huge and unlikely, btw). Doctors’ Services is slowly reaccelerating a little. Hospital Services continues to have problems getting sufficient sample.
• Overall, Medical Care was up 0.1% m/m, but that’s after the continuing ‘insurance’ drag. Y/Y it was at 3.96%, down from 4.15% but looking like it’s leveling out.
• The median category in the Median CPI will be Food Away from Home, +4.63% annualized monthly number. And the y/y Median will decline very slightly again. Was 7.00% in Oct, 6.98% in Nov, 6.93% in Dec. But heading down.
• Biggest upward m/m movements in core categories were in Jewelry/Watches (+48% annualized monthly), Mens/Boys Apparel (+22%), Lodging Away from Home (+20%), Motor Vehicle Maint/Repair (+13%), and South Urban OER.
• • Biggest decliners were energy things, including Public Transportation, plus Used Cars (-27% annualized monthly figure), and Car/Truck Rental (-18%).
• Core ex-shelter: this includes core goods decelerating rapidly and core services accelerating so perhaps isn’t as useful as sometimes: 4.48% y/y, down from 5.2% last month and the lowest since April 2021. But if it stayed there, then it’s hard to get core to 2%.

• While I’m waiting for the diffusion stuff to calculate, a word on what this does to the Fed: nothing. The Fed is aiming for 5% and then will keep rates high for a while unless something breaks.
• Do markets love this data today because it means they were worried about a more-hawkish Fed, with higher rates or higher-for-longer? Or do they think it means the Fed will in fact start easing this year as the curves impound?
• In my view, the latter is really unlikely. I can see the Fed starting QE again if auctions start getting difficult, but in my view there’s no evidence here that we’re going right back to 2% inflation and the Fed has been loudly consistent about this.
• To be sure, they can turn on a dime and they have previously, but…I just think market pricing is really optimistic.
• This [chart below] is consistent with the good news from Median – for the first time, our diffusion index has declined smartly. It’s still above the highs of the last couple of inflation ‘spikes’ (which no longer look like spikes!), but moderating.

• This chart is not quite as good. The mean CPI is falling more because some high outliers (cars e.g.) are coming back to the pack, and some are moving from low to the low tail, and less because the middle is shifting a lot. Look at how >5% is barely declining.

• I mean, that’s not TERRIBLE news, but obviously we need to see the “<2%” get close to 50% if the Fed is going to be confident they’re back near their inflation target. • One more point and then I’ll prep for the call. A lot of the positive-news things are well along towards delivering what they’re going to deliver. Health ins won’t be a drag in 2024. Used cars won’t drop another 20%. And >>
• >>the dollar has turned south so core goods won’t be in retreat forever. The case for inflation going back to 2% rests on rents turning, and on wages slackening. And while those are expected, there are scant signs of them yet. So hold off on the celebrations in the Eccles bldg.
• OK, let’s wrap up and get to the call. Thanks for subscribing. at 9:35ET I’ll be on this call; join if you want to hear me say what I just tweeted. 🙂 [NUMBER REDACTED]

The CPI figure was broadly in line with expectations, which means it was a “something for everybody” kind of number. Disinflationists see continued broad progress towards the Fed’s 2% PCE target, while sticky-inflation folks see the rents and core-services numbers and shake their heads, tsking ominously.

Two broad observations:

First, the disinflation from core goods is ‘on schedule,’ with Used Cars and other core goods categories doing approximately what they are expected to do. But the problem is that core goods inflation is down to 2.1%. If you are looking for the whole number to go back to what it was pre-COVID, you need core goods in mild deflation and core services down to 3%. But both parts of that story are difficult. With the world de-globalizing and near-shoring, it is going to be difficult to see core goods back in an extended period of mild deflation. Probably 0-1% is the best we can really hope for. And that means that the core goods sponge has been mostly wrung out. And core services back to 3%, even if rents are actually peaking (and just not showing up in CPI yet)? Well, core services-ex-rents remain pretty buoyant. So how do we get that back to 3%?

Second. The interesting part of the story is coming up. Inflation is probably returning to “the mean,” but what is the mean inflation now? For a quarter-century it was stable at 2-2.5%, but prior to that it had never been very stable. There are feedback loops in inflation, and those appear visibly to be at work here: higher wages help support higher services inflation, and rents, which in turn support higher wages. Social Security and other wage agreements that are explicitly linked to inflation help this process along. But it means this: the mean is not stationary. The real question of 2023, and probably 2024, is this: what is the mean, now?

My guess? It’s 4%ish, or even slightly higher. It’s very unlikely to still be 2-2.5%. Ergo, it is going to be very hard for the Fed to end 2023 in a happy mood…which means that it is going to be hard for investors to end 2023 in a happy mood!