March headline and core CPI rose 0.1% and 0.4% respectively, month over month, generally in-line with expectations of 0.2% and 0.4%, respectively. Recall "core" excludes the volatile food and energy categories. While index level CPI figures generally met expectations, under the hood, there's a lot more to like. At least for this month, the inflation glass is looking half full again, with March looking a lot more like the 4th quarter of last year than January and February of this year.
As a reminder, the Fed is looking at inflation in 3 buckets now: goods, shelter, and services ex. housing (this last bucket is often referred to as "core" services). This month, the components of the 0.4% core monthly increase consisted of a 5 bps contribution from goods, 26 bps from shelter, and 9 bps from core services. With shelter still expected to roll over later this year (recall the CPI calculates shelter inflation with a significant lag to what's happening in the market, and both home prices and rent prices have been flat to down for most of the last six months), and core goods and services behaving more encouragingly again, there's reason to hope that we could see more CPI releases that look like today's rather than what we saw in January and February.
Here are the historical charts of the above three buckets, and their respective contributions to core monthly inflation over time.
Thus, if you "mark-to-market" shelter inflation, and take us back to pre-COVID inflation there (let's call it 10 bps a month), with goods inflating 5 and core services inflating 9, that produces 24 bps, or 0.24% core monthly inflation. If you annualize that, we're not far off from our 2% target. As we've seen in recent months, this will most likely be bumpy, and I'll likely be writing at least a few more "two steps forward, one step back" posts before all is said and done. But with the vast majority of economic data points slowing considerably over the last month, that should further help inflation too. Slowing wage growth and slowing job growth should be another tailwind in that regard.
Used car prices have been in focus the last two years, and for good reason. Year-to-date, used car prices have moved up, almost every two weeks. This has been broadly driven not by more robust consumer demand, but because auto manufacturers still cannot seem (or in some cases do not want to) try and get back to pre-pandemic production levels. Less supply against the same demand drives prices higher. In the CPI, however, used car prices have still been negative year-to-date, diverging with the Manheim Index. Most market participants have been expecting used car prices to contribute positively (to be clear, I mean here this category would make overall inflation higher than it otherwise would be, whereas the actual results have seen the used car category reduce overall inflation). Importantly, however, we should highlight that used car price indices like Manheim track wholesale used car prices, not retail used car prices. The former is the price that dealers themselves pay to acquire cars. The latter is what you and I pay the dealers for the cars. The CPI tracks the latter, not the former. One of the unique aspects about the last two years in the car market has been that dealers and other sellers of cars have generally seen higher than normal profit margins. Perhaps some of the recent incongruence in these datapoints suggests used car profit margins are finally normalizing a bit. I'm a little skeptical of that (though I do think it will eventually happen, just not right now), and I still think we'll see used car prices become a problem again for monthly CPI readings. It could also just be a bigger lag in the data, which will make upcoming months look worse than we might otherwise expect too. But we'll see.
Thankfully though, we can look at goods inflation without the impact of cars, both new and used. Below is a chart that does just that. It shows the average monthly increase, seasonally adjusted, for goods categories both including and excluding the impact of new and used cars. Excluding cars is the blue bars. This shows goods inflation is generally receding again after some large spikes to start the year.
Additionally, we've talked about in prior months Medical Care being a distortionary tailwind to Core Services Ex. Rent (so helpful to producing lower inflation). Services Ex. Rent Ex. Medical Care, though still a bit elevated, moved back down this month, which was encouraging, though Medical Care was still a better-than-normal tailwind.
We can also do the same thing for core services that we did for goods above, which is look at the average monthly increase to try and assess inflationary breadth. If more categories are increasing at smaller rates (this is what we call disinflation), but overall services inflation still seems high, we know that it's likely because of a few categories having outsized impacts. The opposite can be true too. If you look at both the average monthly changes (again, this is seasonally adjusted) within Core Services Ex. Rent, however, you do see nice improvement here as well. This chart too shows March looking more like 4Q22 than January and February of 2023.
Here are two other charts that we usually show, which show a slightly less optimistic reading of today's data. First, we show 3 month average inflation, annualized, both at the index level and then showing the median and average price increases for goods and services within our Sub-Indent 4 (recall this is a basket of 55 goods and services categories that make up about 98% of the core CPI). The 3 month average chart is probably the worst looking chart we can produce today, and while the aggregate core figures don't deviate as much as the internals so far this year (January's 0.4% core increase was actually the same as March's 0.4% increase, but the inflationary breadth in January was considerably worse than March), this chart will pick up one bad month (January), one iffy month (February) and one more encouraging month (March). More months like March will reduce the inflationary breadth shown in the orange and blue lines, and bring them back down to levels seen later last year.
Now here is that chart on a monthly basis, and not annualized. While we highlighted that the average monthly increases for goods and services within Sub-Indent 4 ticked down again (blue line in the chart below), the median actually ticked up modestly (orange line), from 0.2% to 0.3%. The median can often be a more reliable indicator than the average, so we'll need to watch this going forward. Thankfully, March's 0.3% median category increase on an annualized basis still shows core inflation a lot closer to the Fed's 2% target than the year-over-year figures in today's release (~3.7% vs. the 5% year-over-year number for both headline and core in the press release) tell us.
Bottom line, today's CPI release had more positives than negatives, and puts us back on the right track down to lower (2%) inflation.
Comments