Archive
How Many ‘Shortage’ Anecdotes Equal Data?
There is a growing list of categories of prices which are seeing abnormal price pressures. At least, they are abnormal by the standards of the last quarter-century! A couple of months ago, in “The Risk of Confusing Inflation Frames,” I wrote about some of the effects we might soon be seeing, and of the risk that some of the known-but-temporary effects will obfuscate more serious underlying issues.
In April, we will get the CPI for March; this will be the first CPI release to have ridiculously easy comparisons against the year-ago month. March 2020 was -0.2% on core CPI, and I suspect the consensus estimate for March 2021 will be something like +0.2%; this implies the y/y core inflation number will jump from 1.3% to around 1.7%, depending on rounding. But as I said, that disguises some of the important underlying pressures that may also start to appear with this number. There is an old saying that the plural of “anecdote” isn’t “data,” but eventually there must be a crossover point where the preponderance of independent anecdotes begins to approach the informational value of data, right? Well, here is a short list of some recent anecdotes and reports of shortages.
There has become an acute shortage of semiconductor chips, which has impacted automobile production (and will that increase prices for what is available?). There is a shortage of shipping containers, causing widespread increases in freight costs affecting a wide variety of goods. Packaging materials, which are also a part of the price of a great many goods, are also shooting higher in price. Worker shortages at various skill levels were reported in the most-recent Beige Book. There is a shortage of Uber and Lyft drivers.
There are other effects that have shown up but I misapprehended the significance of them at the time. Apparel prices have risen at an annualized 9% pace over the last four months. I’d attributed that to shipping, but there is more to it than that. In January US Customs issued a Withhold/Release Order (WRO) on cotton and tomato products coming from the Xinjiang region of China, where forced labor is employed; the order calls for the stoppage of freight with any amount of cotton (or tomatoes, but there is not much tomato in apparel) that originates from that region – even if it is only the thread on the hem. While this and the other effects on apparel are probably temporary, we don’t really know how temporary.
Importantly, we should add to these shortages a growing shortage of housing. The inventory of homes available for sale just hit an all-time low (the National Association of Realtors started keeping track in 1982).
And, as a result, the increase in the median sales price of existing homes just reached an all-time high spread over core CPI (home price increases sometimes have been higher, though it is unusual. For example, in May 1979 the year-over-year increase in the median home price was 16.9%. But core inflation was 9.4% at the time, so the real increase in home prices was only 7.5%).
I have written elsewhere about the fact that there is large divergence right now between what the BLS indicates the effective inflation in the cost of housing is, and what a measurement of asking rents suggest it should be. The significant chart is reproduced below – and the short story is that the divergence dates to the imposition of the COVID-related eviction moratorium. This has decreased the amount of rent that landlords actually expect to receive on average, which lowers effective rents even though every other measure of the true (free market) cost of shelter would be, is ratcheting higher at rates seldom if ever seen before.
Now, this moratorium was due to expire at the end of March, but the CDC just extended it until June (which may be one reason that TIPS breakevens have hit some minor resistance). That’s a little unfortunate since it means that the moratorium will expire right about the time that the CPI is enjoying favorable comparisons versus 2020. The understating of rent and owners’-equivalent rent inflation, since those are a huge portion of the consumption basket, has an outsized effect on CPI. I want to be fair here to the BLS: in an important sense, the CPI data on rents is not wrong because in fact if a tenant pays less because of the moratorium, then that tenant’s cost of living really did go down. Even though in a free market without such a moratorium his cost of living would have been higher, that’s not the question the BLS is trying to answer. The cost of living is lower in such a case. Of course, that’s temporary, and so when the moratorium is lifted we can expect the BLS will also faithfully report the catch-up. Which means that in the summer, when we would have expected y/y CPI to start to decline again as it faces more difficult comparisons to 2020…it may not, because rents will start to catch up. That’s going to toast the marshmallows of a lot of investors.
Now, there’s one more facet of the cost-of-shelter question and that’s whether home prices have risen too far, too fast and so it’s home prices and asking rents that will have to decline, rather than effective rents re-accelerating. This is a reasonable question. It is true that the ratio of home prices relative to incomes is getting back to levels that in the late 2000s indicated a bubble was getting ready to pop (see chart). For many, many years median home prices relative to median incomes was fairly stable at around 3.4x. Some increase makes sense since homes have been getting bigger, but it does give the appearance of being overextended.
However, last week in Money Illusion and Boiling Frogs I argued that the nominal value of certain real assets might be usefully compared to the level of the money supply as a way of assessing their real value. Comparing the equity market to M2 made the former look less frothy, and the argument is that maybe equity investors aren’t suffering from “money illusion” in the same way that consumers might be (so far). But the same cannot be said for the housing market. The chart below (Source: Bloomberg) divides the home price index (from the FHFA) by M2. While home prices relative to incomes look high, home prices relative to the stock of money look quite low. It is interesting how the QE of the early 2010s shows up as a one-time shift in this ratio, followed by a period of stability, isn’t it? It suggests that maybe home prices didn’t fully adjust to the new money-stock reality after the bubble’s burst in 2008 and the subsequent QE. And maybe such a one-time shift happens again now.
But it might also be the case that the current rapid escalation of home prices is the market’s attempt to get the real value of the housing stock to reflect the rapidly increasing value of the money stock. If that’s the case, then it also suggests that median wages probably will eventually follow. The last people to respond to money illusion generally are the people selling their labor.
I don’t know if this is the ‘right’ answer, and my purpose in these articles isn’t to give the ‘right’ answer. I just want to ask the right questions…and I feel like these are the right questions.
Money Illusion and Boiling Frogs
“Twice a day we are all forced to await the quotation of the Zurich bourse. Every fresh drop in its value [of Austrian kronen to Swiss franc) is followed by a wave of rising prices … The confidence of Austrian citizens in the currency administration of the State is shaken to its foundation. The State which is perpetually printing new banknotes deceives us with the face value … A housewife who has had no experience of the horrors of currency depreciation has no idea what a blessing stable money is, and how glorious it is to be able to buy with the note in one’s purse the article one had intended to buy at the price one had intended to pay.” – account of Frau Eisenmenger, recounted in When Money Dies (Adam Fergusson).
“Speculation on the stock exchange has spread to all ranks of the population and shares rise like air balloons to limitless heights … My banker congratulates me on every new rise, but he does not dispel the secret uneasiness which my growing wealth arouses in me … it already amounts to millions.” – Ibid.
These two passages come from the contemporaneous observations of an Austrian living through the early stages of the hyperinflation that followed WWI in that country. I don’t for a minute mean to suggest that the global economies are on the verge of hyperinflation, but I present these as an apt illustration of a concept called money illusion. In the first passage, the writer makes plain that the kronen is buying less and less, in terms of real goods, every day. Similarly, it buys less and less in terms of equity shares. The former, we tend to regard as a negative, and the latter as a positive, even though they are both related in this case to the same phenomenon: the unit of measurement is losing its value, so that it buys less real stuff as time passes. Isn’t that interesting? For someone who is continually investing in the equity market – I’m looking at you, millennials – higher prices should strike us as a bad thing just as higher car prices strike us as a bad thing.
I don’t mention that, though, to suggest that equities are a great place to hide out from inflation. In fact, they’re a pretty lousy place: as inflation rises the multiple paid on earnings declines so that even if nominal earnings are rising with inflation equity market prices can’t keep up. That’s not as bad as holding paper money and watching it go to zero, but it ends up being about the same when the inflation gets serious enough that the market itself collapses – as it did in each example of monetary hyperinflation (Germany, Austria, Zimbabwe, etc) that we have seen to date. But again, it isn’t my purpose today to warn about the dangers of treating equities like real assets when multiples are at nosebleed highs.
The interesting part is the money illusion. The writer in the passages above is uneasy, because while she is making millions she understands that those millions are losing value almost as fast (and ultimately, faster) than she can make them. But for a while the higher and higher prints of the market, the rising value of one’s home, and the accelerating increase in wages makes people feel wealthier. And wealthier people are happier and tend to spend more of the marginal wealth, when that wealth is real. But in this case the wealth is an illusion, because that additional wealth buys (at best) the same amount it did previously.
In classical economics, we would call spending more in this circumstance – despite having a similar claim to wealth in real terms – irrational. Although we use dollars to translate our labor into the things we want to buy, we all understand that we are really trading our labor for those things – it’s just that we need a medium of exchange because no one wants to directly exchange groceries for inflation-focused asset management services. More’s the pity. So homo economicus would regard his increasing millions in the market and not feel any wealthier as he knows the units of account are growing weaker. The money dropped into his bank account through a universal direct stimulus also wouldn’t be treated as actual wealth, since if we handed everyone a trillion dollars then obviously we all wouldn’t be living like trillionaires because the people who sell goods and services would adjust their prices (if they did not, then those vendors are voluntarily decreasing their own claim to the real wealth, by accepting smaller real payments in return for the same amount of goods). Wealth is just a claim on the national product. If everybody’s nominal wealth rises, but the nation is not able to produce more units of real output, then in aggregate we clearly are not wealthier because the pie is the same size. (Now, if you hand everyone a trillion dollars except for one guy, then that guy is poorer and everyone else slightly richer. Ergo, direct cash payments to the poor are clearly a way to distribute actual wealth, especially if those who don’t receive those payments also face higher taxes. So fiscal policy here definitely shuffles the deck of the wealthy. It just doesn’t make us wealthier in aggregate.)
The question of how people behave when they see additional income that comes from a greater money supply, rather than from additional productivity/output, is crucially important in monetarism. In the quantity equation of exchange, MV≡PQ, an increase in the quantity of money and in the velocity of money (MV), which is the total nominal amount of expenditures, necessarily equals the real output times the price level of that output (PQ). The amount that is spent equals the amount that is bought. But how the right side divides between P and Q is very, very important. If there is no money illusion, then an increase in the quantity of money will primarily increase prices while output will remain stable. Shopkeepers are unwilling to part with their wares for a smaller piece of the pie in real terms. On the other hand, if money illusion is rife then producers respond to consumers flush with cash by providing as many goods and services as they can; they view the masses as having more actual wealth to spend and so output increases and prices don’t rise as much.
Unfortunately, it seems that money illusion operates primarily when the quantities involved are small, or narrowly distributed. When incremental money creation is widely distributed and significant in size, then (as the second quote at the start of this article suggests) consumers, suppliers, and investors eventually figure it out. When that happens, a change in M is almost fully reflected in a change in P, as over time it usually is anyway. So the secret of recovering from a negative economic shock by expansionary monetary policy is to boil the frog slowly.
No one involved in current policy circles is interested in boiling the frog slowly. And that means it’s not going to end well.
In this context, the current bubbly stock market looks decidedly better. The chart below shows the S&P 500 divided by M2 (and multiplied by 100 because sometimes I don’t like looking at decimals on my y-axis). Now, the S&P 500 level isn’t the purest look at the total value of the equity market, but you get the general idea here – stocks have outrun the growth rate in the money supply, even over the last year, but the new records we are hitting are mostly on money vapor.
Summary of My Post-CPI Tweets (March 2021)
Below is a summary of my post-CPI tweets. You can (and should!) follow me @inflation_guy. Or, sign up for email updates to my occasional articles here. Investors, issuers and risk managers with interests in this area be sure to stop by Enduring Investments!
- Another #CPI morning as spring is getting ready to spring here in the northeast. And with spring, more activity.
- Today’s #inflation figure will be the last one to be compared to pre-COVID year-agos. The easy comps start next month. So, while consensus today is for +0.2% on core, that will still not move core much since last Feb it was +0.21%.
- Headline will jump a bit, because gasoline has been rising, but the real headlines if any will be below the hood. Last month core was flat, due to soft housing and a somewhat surprising decline in used cars and pharmaceuticals.
- I expect we will see the used cars number reverse this month (Black Book was strong), and with the end of lockdowns might see some strength in the covid categories. Pharma price hikes ought to eventually show up. But they’ve been confounding me.
- Global supply chains are a disaster and raw materials and packaging prices are spiking – the Texas freeze shut-down did NOT help polypropylene prices! – so goods prices ought to continue to rise. Eye on apparel as the canary there.
- The rent story is a passing one. But probably not yet, which means OER and Primary Rents should still look a little soggy. Asking rents are jumping, but measured rents are not – because of the eviction moratorium. If you’re not paying rent, it’s not a cost of living!!
- In this quarterly chart, you can see the divergence between asking and realized rents. The divergence began in Q3…which is when the eviction moratorium was enacted. That’s not coincidental.
- In the more-recent COVID relief bill, the eviction moratorium was not extended past March 31st, which was a bit of a surprise. That could still change, but when the moratorium eventually expires I am pretty sure we’ll see a rapid catch-up of rents. But not today’s story yet.
- My calculations are that if the end of the eviction moratorium caused effective rents to catch up to asking rents, the effect on OER and Primary Rents would add something like 0.9% to core CPI. (!) So that’s 2021’s following wind to prices.
- As always, I want to be sure to remind you that the Fed does not care about inflation any more. Someday they will, but not yet. They’ve even stopped reporting weekly M2! They believe they have the tools to stop inflation so they’re not worried. Ergo, you’re on your own.
- Although not exactly. We’re here to help. If you have interest in how to hedge/invest in the inflationary period approaching, visit https://www.enduringinvestments.com
- And for a summary of today’s series of tweets, you can check later at https://mikeashton.wordpress.com
- Thanks for coming with me on this #CPI journey this morning. Buckle in.
- Core looks like 0.10% flat, dropping y/y to 1.28%. That will be the low for…probably 10 years.
- Even the +0.1% was higher than the last couple of months. But the easy comps start next month. Then some hard ones. But by the time we get to the hard comps, rents should be catching back up. Worth looking at my “conflicting frames” piece on the blog.
- Used Cars and trucks surprisingly fell again, -0.91% m/m. That’s at odds with the private surveys, but there’s sometimes a wiggle before they catch up.
- Apparel dropped -0.74% m/m. Also odd. Pharma PLUNGED -0.75% m/m. As a result, the core goods y/y figure dropped to 1.3% from 1.7%! That’s the story.
- Rents were actually okay! OER was +0.27% and Primary Rents +0.20%. Really, today’s story in some ways is the opposite of what I expected. Rents solid, goods prices soft.
- Really odd stuff today. Airfares -5.1% m/m, despite the heaviest air traffic in a year. Lodging Away from Home -2.3%, ditto. Although some of that might be a weather effect. Still anti-anecdotal.
- Doctor’s services, though, jumped +2.01% m/m! To 5.1% y/y, highest in a while.
- Hospital services remained a bit soft. Still, the overall Medical Care subcategory accelerated slightly to 2.00% y/y from 1.95% y/y, despite the drop in Pharma, thanks to docs.
- So Core ex-housing dropped, to 1.16% from 1.25%. But for weird reasons, not services but goods!
- Core goods and services back together after core goods had rocketed ahead. Again, this is super weird. Shipping costs are through the roof. Packaging and raw materials prices are having moves like we’ve never seen. And goods prices are declining?
- sorry…decelerating. Let me be precise.
- Biggest monthly declines in core: Jewelry (-29.9% annualized), Lodging Away from Home (-24.4%), Public Transportation (?) (-24.0%), Men’s/Boys’ Apparel -23.6%, Infants’/Toddlers’ Apparel -21.6%, Tenants’/Household Insurance -13.5%, Used Cars/Trucks -10.3%.
- The decline in insurance makes sense. Insurance companies are having to give rebates because their loss ratios were too good (that is, they didn’t have to pay out as much as underwriting had expected). So that will be a CPI decline. I get that. Only 0.4% of CPI though.
- Largest core increases: Car/Truck Rental (+135% annualized), and that’s it for >10% annualized.
- Pharma y/y. Difficult supply chain for APIs and more announced price hikes in January than is normal. And prices are falling. That’s a conundrum.
- This is the divergence I mentioned in private surveys vs CPI for used cars and trucks. I could believe it cresting at a lower level but the latest zig higher suggested we have another zag in CPI. Next month, maybe.
- Brilliant catch. Although lockdowns are lifting, the BLS had trouble doing some collections. More likely weather, but blamed on Covid because we blame everything on Covid. But that explains a lot of weird moves. Good catch @TOzgokmen
- Well, that makes next month even more interesting. Because the rent numbers aren’t likely to have been much impacted by weather, but physical goods prices?
- And again, like I said the other surprise was that rents did not continue their recent trend of softness. Y/y on OER was flat. There’s a lot of catch-up ahead.
- One thing which has changed in the last year which isn’t likely to change back very soon is the VOLATILITY in the monthly CPI figures – not just the core or headline, but the subcomponents. That will persist for a while.
- OK, four pieces charts. Food and Energy relatively normal.
- Core goods – slightly off the boil but will be interesting for sure to see how much of this was due to “collection problems” at BLS.
- This is where there’s real weakness, but it’s airfares, hospital services…though doctor’s services added a bit this month. Insurance rebates likely pressured this in February. Again lots of volatility.
- And then Rent of Shelter, normally the least volatile. This month, rents were actually pretty normal but lodging away from home dragged further. I do think that hotel rates are unlikely to keep sliding. This might also be a collection issue.
- Think that’s going to do it for today. The bottom line is that rents were stronger-than-expected, lots of other things weaker, but collection issues make it easy to be skeptical that goods prices are suddenly decelerating. Next month we’ll get hopefully a cleaner picture. BUT…
- …BUT we will also get the beginning of the severe base effects. In three months, core CPI will be near or above 3%. We all know it, but it will be interesting to see if markets get nervous anyway. Thanks for tuning in. Stop by https://www.enduringinvestments.com !
- FWIW, looks to me like median CPI ought to be more like +0.26% compared to core +0.10%. My confidence in that is lower than usual because of something quirky with my spreadsheet, but it highlights that the core was dragged down by large declines from small categories.
The best observation of the day wasn’t mine. The problem with pulling in data automatically, rather than reading the report, is you will miss the footnotes! And the footnote pointed out by @TOzgokmen was a very important one. It slaps much larger-than-normal error bars around what are already more volatile-than-normal data. I suspect that this was not likely to have a big effect on rent data collection, but more likely on goods and services where specific outlets were likely to be closed by the bad weather. I may be wrong about that, but it does go a long way to helping understand the weird fact that rents were solid (instead of weak as expected) and goods prices were weak (instead of strong as expected). One never should put too much weight on any one month’s figure, but this diminishes the anecdotal value even more. On to next month.
