We’ll dedicate today’s “Data” section to the Consumer Price Index report out this morning. Ordinarily it would only merit a piece of this section, but with so much investor attention on the topic of inflation we need the space to discuss it completely.
#1: We will start with how today’s print fits into the historical record. The chart below shows CPI headline (black line) and core (red line) inflation back to 1970. Today’s 5.0 pct headline and 3.8 pct core inflation are the highest since September 2008 and May 1992, respectively.
Since the whole “highest inflation since X” is a popular headline today, let’s see why headline/core inflation was high in 2008/1992 to see if these analogies really work:
- 2008 was when oil prices hit their all-time highs, at $140/barrel in June. Even in September, they were still +$100/barrel. This drove gasoline prices +32 percent higher, responsible for a third of September 2008’s 5.0 pct headline inflation.
May 2021 is very comparable to that: gasoline inflation of 56 percent is fully 40 pct of this month’s 5.0 percent headline inflation rate.
- May 1992’s core inflation print of 3.9 percent was actually right in the middle of a structural move lower in non-food and energy prices. They fell from +5 percent in 1991 to 2.6 pct at the end of 1994 and, as the chart shows, remained low thereafter.
Unlike May 1992, May 2021 shows several specific areas of core inflation, which we will now address.
#2: May 2021 headline and core inflation numbers are both quite lumpy, with specific items driving the overall results.
Starting with the driver of the 5.0 percent headline inflation number, which is noticeably higher than 3.8 pct core inflation:
- Energy is 6.9 percent of the Consumer Price Index and showed 28.5 percent inflation versus May 2020. This means Energy prices were 1.8 points of the 5.0 percent headline inflation rate, or 36 percent of that number.
- Food is 13.9 percent of headline CPI, and with a 2.2 percent annual increase this category actually shows both low and declining inflation.
For some long-term context around Food inflation, which rightfully gets a lot of attention, here is a chart back to 1970. As you can see, it has been in a band between zero and 5 percent since 1990 and current levels sit right in the middle of that band. We’ve highlighted the peak back in 1974 (20 percent!) just as a reminder that it wasn’t just Energy prices that caused the early 1970s inflation spike. Food was 25 percent of the CPI basket back then, versus 14 pct now. When you hear commentators say “we’re going back to 1970s inflation levels”, just keep that in mind.
Moving on to the May 2021 core inflation rate of 3.8 percent:
- Used car prices are almost a quarter (23 percent) of that increase. Yes, you can sort-of blame fiscal and monetary policy for giving consumers the capital for a down payment on a vehicle financed with low interest rates. But the chip shortage that is curtailing new vehicle production is just as much to blame for rising 1–3-year-old used car prices. The Manheim auto auction data shows that is where the bulk of rising used vehicle prices actually sits.
- Transportation services inflation represented 15 percent of core inflation this month even though it is only 6.6 pct of this index. The key drivers there were motor vehicle insurance (+16.9 percent inflation) and airfares (+24.1 pct). Both were significantly lower last year and are now rebounding.
- Excluding these 2 categories, core inflation last month was 2.3 percent, very close to the last pre-pandemic print of 2.4 percent in February 2020.
#3: Looking forward, the US inflation story will hinge on the “Shelter” category, with Owners’ Equivalent Rent and Rent of Primary Residence making up 24 percent and 8 percent of headline inflation respectively.
The trends here do seem to have bottomed at around 2 percent annual inflation and are starting to turn higher, as this chart back to 2000 shows (OER in blue, Rent in Red, and CPI in black). What’s really notable about this data, however, is that Shelter inflation ran well ahead of overall CPI inflation for almost all of the last decade (2012 – 2020). Therefore, even if Shelter inflation reverts to form over the next 1-2 years, we’d need to see non-Shelter inflation reset to higher levels in order to see structurally higher inflation.
Summary: dig through the CPI report as we’ve just done and you’ll see that there’s very little to support the argument of structurally higher US inflation. Yes, this is the same message as we delivered when we tore apart last month’s report. We doubt June’s data will be much different.
Investment implications: today’s CPI report fits nicely with our Markets piece last night which explained how 10-year Treasury yields look at long run inflation trends rather than just 1-2 years of data. Yields today continued their retreat from their March highs, lifting the S&P to a new all-time record. We’ve been a touch cautious (but not bearish) on US stocks lately, and a modest new high alone is not enough to shift our view. Clearly, a decent chunk of our “no secular inflation” thesis is already priced into Treasuries. Big Tech should see a small catch-up rally as a result. But as for the next move higher in large caps, we still think that will only happen as companies report Q2 in July and signal their outlook for the rest of 2021.