A slew of inflation data came out over the last several days, highlighted by the August CPI data on Tuesday. Other notable releases included August’s Producer Price Index (which tracks prices companies are receiving from vendors and suppliers) on Wednesday, and Import / Export prices, New York (Empire) Fed and Phila Fed surveys on Thursday. The key takeaway is that CPI was very disappointing, but everything else was pretty encouraging. The summary conclusion in my view is that inflation continues to get better (my best guess is right now we’re at a pace of 4-5% annualized inflation vs. 7-8% in May and June), but perhaps not by as quickly as I thought a week or so ago. I’m going to break down this week’s data in two posts. The first is dedicated to the CPI, and the second will be everything else.
Let’s start with the CPI on Tuesday. Headline CPI (including food and energy) came in at +0.1% vs. an expected decrease of 0.1%. Core, however, which excludes food and energy, was the bigger miss, coming in at +0.6% month-over-month on a seasonally adjusted basis versus expectations for a 0.3% increase. Prior to this release, investors had been hoping August’s core release would look similar to July’s +0.3% increase, which implied annualized inflation of sub-4% (0.3% x 12). Generally speaking, this release was a step in the wrong direction, but not so much that we can’t still conclude that inflation is getting better. It just may not be getting better as fast as some (like me) thought.
I like looking at core CPI data in several ways. First, I should state that I tend to emphasize the seasonally adjusted (SA) month-over-month changes much more than the year-over-year increases, particularly right now. Why? Because as I’ve detailed in other posts, the month-over-month numbers give you a much better indicator of current inflationary momentum, rather than incorporating price changes that happened 6-12 months ago. The press almost always quotes the year-over-year numbers, mostly because they're higher and more headline-grabbing, but this is misleading (that's why you come here!). Investors can compare the logic of looking at year-over-year instead of month-over-month to looking at LTM EBITDA as a proxy for a company’s earnings instead of forward (or next twelve months) EBITDA when the outlook is likely to differ considerably from what happened in the last twelve months. As we all know, things can, and do, change quickly, and the last two years have been clear evidence of that.
Specifically, I like looking at the CPI data in the following ways. First, I like looking at category levels and looking at the average and median increases within those categories. Second, I like looking at the number of categories increasing at various thresholds, which gives me an indication of inflation breadth. Third, given what’s happened over the last few years, I am focusing on the contributions from goods versus services, and then also services with and without rent (which helps captures services inflation outside of housing).
Note that I’ll reference “Sub-Indents” several times here (and which I’ll notate as “SI” going forward). These reference the category depth level provided in Table 2 of the CPI releases (Table 2 gives us the CPI Index by Detailed Expenditure Category). For reference, Sub-Indent 4 (or SI4), comprises 55 different categories of goods and services, and whose weights captured 97% of the core CPI index in August (so almost all of it). Sub-Indent 5 (SI5), by contrast, captures 101 categories but only captures about 77% of the core index. Because SI4 captures essentially the totality of the index, I tend to give greater weight to that, but most of the time, SI’s 4 and 5 tell the same story, and that’s true again in August.
Here's the median and average month-over-month changes for the 55 and 101 categories provided in SI 4 and 5, and then how they compare to historical levels. You can see that we're still elevated versus pre-COVID levels, but that we're seeing improvement from the higher levels from earlier this year.
The next two charts show how SI4 compares to the reported aggregate change for the core index, for perspective. You can see that for a good chunk of time earlier this year, the average was actually worse than the reported aggregate number, whereas now, its better. The same goes for the median increase.
All of these charts indicate that the median and average monthly increase on a category level seems to be hanging in around 0.4%, which is slightly better than the reported aggregate index figures. Said differently, the average and median monthly price increases across the core CPI index has averaged about 0.4% over the last several months, which implies annualized inflation of about 4.8% (0.4% x 12). This is slightly better than the reported core CPI increases of 0.7%, 0.3%, and 0.6% (which averages closer to 6-8% inflation on an annual basis), indicating that several categories are driving outsized impact on the overall index. The charts continue to make clear, however, that inflation remains stubbornly high, but we are moving in the right direction.
Now, here’s a couple of charts showing how many categories are increasing at various thresholds on a monthly basis. Ideally, you'd like fewer categories increasing by greater than 0.3% (which implies annualized inflation of 3.6%). To get back to ~2% core inflation, you need many more categories increasing by closer to 0.2% per month. The overall takeaway from these charts is that inflation isn’t showing a ton of improvement, but it is showing some. Even August’s “miss” shows levels that are still below the worst levels seen earlier this year.
Now, let’s look at a comparison of goods inflation versus services inflation, and try to assess which categories are rising the fastest (or contributing the most). The first two charts show the monthly contributions to our seasonally adjusted core CPI number (which was 0.6% in August, for reference) from goods and services.
What’s important to note here is three things: first, that goods inflation is moving in the right direction from the very elevated levels of last year and earlier this year. Second, despite the improvement, goods inflation is still elevated versus pre-COVID levels. Third, and perhaps most notable, is that goods inflation was essentially zero prior to COVID. This means that almost all of the inflation in the economy prior to COVID was coming from services. This is encouraging in that goods prices still have room to fall, which will put additional downward pressure on overall inflation, but that's only really comforting if services inflation is also moving in the right direction. As we’ll see in a moment, excluding housing (whose impact acts with a lag, but which also appears to have peaked), this does appear to be the case.
Now let’s look at the contribution to inflation from services. We’re going to look at this with rent, the rent contribution itself, and then without rent.
Here is a chart showing the services contribution both including and excluding housing, overlaid on top of each other, for better perspective:
As these charts show, non-housing services inflation remains somewhat elevated versus pre-COVID levels, but it too is moving in the right direction. Total services is being inflated by rent. This month’s services contribution ex. rent of 13 bps compares to the 18 month pre-COVID average of 7 bps, for perspective, but we can see that for a good chunk of 2019, the average was much closer to ~10 bps. Thus, there’s an argument to be made that even August’s “hot” CPI figure was not too far off from pre-COVID services inflation levels, again, excluding housing.
The overall conclusion from this month's CPI report is that August’s CPI was a step backward on the path to lower inflation, but we’re still on the path. More on this in Part 2 of this post.
For those that are interested, below are charts with contributions from each key goods and services category to our core m/m SA CPI figure over time to provide additional perspective. The takeaways are that several goods categories, especially household furnishings, automobile related, and recreational goods continue to contribute well above their pre-COVID levels. Within Services, most categories are showing nice improvement, but rent and medical services stand out as stubbornly inflationary categories.
Comments