Polymarket Prediction: Will inflation be 0.4% or more from April to May?
Inflation is in the CPI of the beholder
It is with great humility that I am wading into the world of prediction markets by making my first forecast! I say “with great humility” because I am terrified of getting it wrong and looking foolish: predicting headline inflation from month-to-month is basically a crapshoot due to the sheer amount of noise involved. As Kenneth Galbraith would say, “The only function of economic forecasting is to make astrology look respectable.” That being said, part of being a gambler involves being vulnerable and going out on a limb — it’s not gambling unless it hurts when you lose! — so here goes nothing.
The specific prediction market I will be forecasting is linked here; at the time of this writing, one share of “Yes” is currently trading at 70 cents, meaning the crowdsourced prediction from those with skin in the game (shoutout to NNT!) is pricing in a 70% chance that headline inflation from April to May will rise by more than 0.4%. If you wager $70 today on “Yes”, you stand to win $100, for a tidy 43% return.
Easier said than done. For proof of just how impossible it is to precisely predict headline CPI, look no further than last month’s aggregate Bloomberg prediction in which 50 of the most serious, competent, and trusted bankers in the world were off by 5+ standard deviations***! Look at where this yellow dot spawned:
***Strictly speaking, this standard deviation comparison is meaningless since it's comparing a point to the standard deviation of a different distribution. The variance of estimates of the mean is different than the variance of the underlying distribution. I just wanted to highlight how dramatically wrong their forecasts were.
When this yellow dot appeared last month, every “dove” in the world simultaneously threw up their hands and parroted the same mot-du-jour: transitory! This is transitory inflation! This transitory word still dominates the discourse, but is also easy to hand-wave away as a nonconcern in our prediction market context. That is because in Fed-speak, “transitory” means "this year and next year”, so for predicting inflation in the ultra short-term (May 2021) it really doesn’t matter if inflation is a short transitory burst, or endemic. It only matters that it is currently bursting — which we all agree it is. This week, even Janet Yellen admitted that "I expect [high inflation] to last for several more months … through the end of this year".
The next most popular dovish talking point is that the current CPI numbers suffer from depressed base effects. This is an artifact of how the CPI is constructed: it is measured in YoY terms, and the chaotic start of the pandemic was happening 12 months ago, thus leading to wonky numbers. At first glance, this is compelling — of course those depressed base April 2020 numbers would feed forward into artificially inflated April 2021 numbers!
The most succinct refutation of this “reflation” argument is by Hoover Institution economist John Cochrane who rebuts it in a recent blog post. He explains that yes, if you start from a low base, you can see a lot of growth, and yes, drawing a line from last May to today shows an unusually higher slope because last May was unusually low. But upon visual inspection it is evident that an uptick is clearly visible at the rightmost tail of this graph… so to the extent that we are witnessing a return of prices to normal after a COVID blip, the accelerating right tail of this graph suggests that was done and over with last summer.
Even if this base effect excuse did hold water — which it may, or may not — Harvard Business School professor Alberto Cavallo (the world’s leading expert on how COVID impacts CPI measurement) estimates that it increases the annual inflation rate by only one percentage point. So it would still only be a small part of the story. Not to mention that our Polymarket wager only concerns itself with month-to-month inflation, so the year-over-year base effects are rendered moot.
The third point that a dove would then regurgitate is pointing to the existence of extraordinary one-off items in last month’s inflation report. Gas prices went up by a lot and hence are responsible for April inflation, they say! Right? But those kept rising in May, too, blowing past their April highs… so, in attempting to forecast April to May inflation, we can conclude that gas prices won’t provide any soothing deflationary pressure.
Next, the dove would opine about how used car prices were up 21% in April. In fact, media outlets have aggressively latched onto this convenient “used cars” narrative for whatever reason… perhaps it makes for grabby headlines. Yet, it can be summarily dismissed simply by breaking CPI down into its components, isolating used cars among them, and looking at the time series evolution. Just by eyeballing it, I can’t foresee this green line slowing down any time soon, let alone magically reverting back to base 100 in one fell May swoop. Even if this green line were to slow down a little, used cars still only account for a meager ~2% of the CPI basket, so constantly pointing to them as an important “one-off” driver of April headline inflation is a cope. No deflationary relief exists here either… only continued inflationary pressure. I expect the used car shortage to keep getting worse, not better.
We now turn from examining measurement quirks to laying out a multitude of fundamental economic factors, which each in turn support a hawkish outlook. Taken together, they paint a picture of a market that is more concerned about inflation right now, in the month of May 2021, than at any other point in my lifetime. We have reached maximum inflationista!
1) Everyone unanimously agrees that price pressures are increasing simply because demand growth is outstripping supply growth (economists call this demand-pull inflation — lots of money chasing few goods). This is due to screwed up supply chains, rapid reopening, pent-up demand, and pockets flush with cash.
2) Rising costs of production. Lumber is the commodity you have probably heard about most prominently, but pretty much all other commodities are booming too — steel prices have tripled, the Dow Jones Commodity Index is up 70%, and the Producer Price Index (PPI) is hitting record highs. To the extent that global commodity prices are passed through to domestic prices, that will automatically create inflation (economists call this cost-push inflation).
3) America is in the middle of a severe national housing shortage. While home prices do not filter directly into official inflation statistics, there is significant upside risk wherein they spillover into mortgage and rent prices (which account for 42% of the CPI basket!). Anecdotal evidence suggests that this is already well underway.
4) The percentage of businesses reporting that they cannot find qualified workers is at a record high. I am usually very unsympathetic to such "worker shortage" complaints, but the expanded UI — for better or worse — has absolutely kept people out of the labor force as a large chunk of them have raised their reservation wage. The point here is that shortage of labour tends to cause inflationary pressure.
5) The U.S. personal consumption expenditure (PCE) core price index is one of the main ways the Fed monitors inflation. It is closely tied to CPI. Two days ago, it surged to its highest level since 1992:
6) The USD is at a 3-year low. The way I look at it — and this is a vast oversimplification — is that a weak USD buys less in foreign goods, which increases the price of imports, contributing to inflationary pressures.
7) Inflation fears are skyrocketing in corporate America, as quantified by Bank of America, who tracks the number of mentions of inflation in Q3 earning calls. Mentions have exploded nearly 800% YoY:
This textual analysis supports the findings of an NBER working paper released just 4 days ago that finds by any typical definition of “anchored” expectations, the inflation expectations of U.S. managers appear far from anchored.
Checking in with the bond market (i.e. where corporate America parks its cash), the five-year breakeven inflation rate just reached its highest level since April 2011. The 10-year breakeven inflation rate just rose to its highest level since March 2013.
8) Not only are corporate inflation expectations unanchored, but in the general “normie” population, you're starting to see a shift in psychology around inflation purchasing behavior as well. As per Warren Buffet, “We are seeing very substantial inflation: We are raising prices. People are raising prices to us, and it’s being accepted.” This shows up in the survey data: last week a New York Fed survey reported that in the upcoming year, consumers are anticipating gasoline prices jumping 9.18%, food prices gaining 5.79%, medical costs surging 9.13%, college education climbing 5.93%, and rent prices increasing 9.49%. A separate survey conducted last week by Shopkick showed that 77% of Americans were aware of an acceleration of the inflation rate. Ditto with a 5-year chart of Google Trends for the word “inflation”, which peaked after last month’s CPI data release:
9) Perhaps the most salient explanation for inflation is simply the trillions of dollars of new money being printed. In the name of expediency I am not going to go into detail on this point; there are thousands of articles and books on it. Put simply, the Fed creates more money —> the value of money decreases. Here is M1 money stock, presented without comment:
10) Directly following from the “brrrrrr” money printer hypothesis is the waning credibility of the Fed. I simply don’t believe that the toothless Fed nor the rudderless Biden administration has the Volker-esque guts to deal with inflation should it continue to spiral out of control. How can they possibly put these inflation expectations back in the bottle? Sassy tweets? They don’t even have the stomach for even a 1% drop in asset prices, as proven a couple weeks ago when Janet Yellen backtracked some pretty milquetoast comments in response to spooking the market.
Flip flopping in this manner is a sign of weakness — if they are panicking over a 1% drop in the market, they surely don’t have the stomach to raise interest rates, end QE infinity, or shrink the balance sheet. I flat out do not believe this administration will ever in a million years take the swift and painful fiscal and monetary actions needed to control inflation. As Stony Brook University economics professor Gabriel Mihalache points out, the next step after “transitory inflation” will be "what's so bad about inflation after all?", followed by "Only a robust burst of inflation can save our overindebted government and middle class." The politicians will simply keep moving the goalposts rather than confront the problem. Book it.
11) This last point is only half-joking: always do the opposite of what Justin Wolfers and (former Enron advisor) Paul Krugman predict. The former suggests that inflation this month is “already slowing”, while the latter calls inflation this month a “nothingburger”.
We got kind of lost in the weeds there, discussing the big picture outlook. So let’s take it back to actionable and pragmatic gambling advice: headline inflation has gone 0.4 —> 0.6 —> 0.8 in the last 3 months, and historically there is decent correlation from month to month. I would be shocked if in May it dropped below the February level. Using my economic intuition — my gut — I would put the chances of it dropping below 0.4 this month at less than 30%, making it a +EV bet.
There are a couple rogue data points that hint at slowing inflation this month, namely PriceStats and the Cleveland Fed Inflation Nowcast. Despite scouring the internet, those are really all I could find to support the dovish case, and they both got it wildly wrong last month. I don’t find them particularly convincing in the face of all the other evidence, and would confidently bet on them undershooting again this month. Although it would be a feather in their cap if they do get it right.
The strongest holistic argument I can come up with for why I might be wrong is that I have a lot of sympathy for the fact that we are living through extraordinary times… this leaves the bands for potential outcomes very wide. Big data surprises are possible, and the macroeconomy can evolve in unexpected ways. Last month’s rise in CPI could easily be a one-time deal, reverting back to normal this month. Furthermore, perhaps it is supremely foolish of me to “time the market” and triumphantly declare that we have entered a new high inflationary regime. THIS IS NOT FINANCIAL ADVICE!
There are only a small handful of existing macroeconomic/inflation markets in the nascent prediction markets industry. Over at Insight Prediction, we are working to innovate and create more of them! Please sign up to be notified for when our website launches by clicking this button: