Jan + Feb. '25 CPI: Ugly in Jan, Tolerable in Feb

This month's post will account for two month's worth of CPI data, since we missed last month. To sum it up, January's data, as has been the case for the last several January's, was pretty discouraging. February's is better--and tolerable--though a bit less encouraging under the hood. Let's go through it.

CPI, both headline and core, increased at a seasonally adjusted +0.2% this month, compared to 0.5% and 0.4% in January. February's release was better than expectations for an increase of closer to 0.3%, while January's was worse (consensus there was also about 0.3%). Unrounded, February's figure was 0.23%, while January's was 0.46%. February's figures resulted in headline CPI rising 2.8% over the last 3 months, while core CPI rose 3.1%. This is an improvement in both regards from January's 3.0% increase in headline CPI year-over-year, and the 3.3% increase in the core. As we've done for months, we'll make the case that these reported figures are overstating inflation in the real economy, predominantly due to shelter, but more on that later. Below is a table summarizing these results and trends.

As we normally do, let's look at these reports through our "three buckets": goods, "core" services ex. shelter (sometimes referred to as "super core"), and then shelter. To save from the confusion of throwing a lot of numbers at you, here's a table showing the trends in each of these three categories:

Ideally we'd like the total estimated month-over-month, seasonally adjusted core CPI to add up to about 0.2% each month, since on an annualized basis this produces inflation of about 2.4%. February's 0.23% puts us right in that range (though January's, which was almost double, reminds us not to overweight one month's results).

You might be asking, isn't that (2.4%) above the Fed's 2% inflation target? Yes, it is. But the reason 0.2% in the CPI is generally tolerable is because the Fed actually measures inflation using another index, called the Personal Consumption Expenditures Index, which has different weights. These different weights have historically resulted in about 30-50 bps difference between core PCE and core CPI on an annualized basis (with PCE being lower than CPI). Consequently, 2.4% in the CPI should get us pretty close to 2% on PCE.

Let's now go through each bucket in more detail. As noted above, goods contributed +5 bps this month, while services ex. shelter contributed +5, and shelter contributed +13. This compares to the pre-COVID averages of 0, +7, and +11 for those categories, respectively. So goods and shelter were higher (again) this month, while core services ex. shelter was lower.

Starting with goods, which as noted above, contributed +5 bps of our +23 bps this month. In contrast to last month (January) when car prices accounted for the bulk of the goods contribution, this month's goods increases were a little broader. Household furnishings, Apparel, Alcoholic Beverages and "Other Goods" all contributed to this month's +5 bps. In a pre-tariff world, we used to be able to write off these goods increases as aberrations that would revert the next month or soon thereafter. Now, it is less clear who's trying to get ahead of tariffs and what's simply a price increase for it's own sake. Given the softness in consumer spending since January, however, goods producers may not be able to pass through all of the tariffs they see (or at least not yet). But that remains to be seen, and unfortunately, our base case can no longer be to simply dismiss goods inflation as exceptions that would quickly mean-revert back to zero. Hopefully we don't live in this new world for long.

Core services ex. shelter ("super core") was one of the highlights of the month, coming in at +5 bps contribution as well. This is well below January's +22 bps contribution, and below the +7 bps pre-COVID average as well. Here though we may need to be a little careful, because airfares (part of transportation services) were a -5 bps contribution all by themselves. Given the dynamics in the airline industry (low cost and ultra low cost carriers being forced to cut capacity, which will likely drive fares to major cities higher in coming months and years), that kind of benefit is not likely to persist, unless of course the consumer is truly rolling over. Car insurance thankfully only contributed +1 bps this month, about in-line with its pre-COVID average (which is zero). Education, recreation and other personal services were positive contributors above their historical averages, however, so there's certainly a lot of moving pieces in this section of the data this month. It's therefore hard to draw any firm conclusions about core services ex. shelter this month, but thankfully, the net net this month was a good number. Hopefully all the puts and takes net out to a similar result in future months.

And lastly, shelter. Shelter has been a thorn in our side for years now, and we discussed the favorable outlook for renters in a recent post here. This remains true today. With the exception of June and September of 2024 (which seemed to randomly have smaller contributions), February's +13 bps was the lowest shelter contribution to monthly core CPI since last October (when shelter also contributed +13 bps), though it is still above the pre-COVID average of +11 bps. As the chart below shows, the general trend remains down for CPI shelter, but the volatility and extremely long lags has kept aggregate inflation way above where it actually is in the real economy.

Below there are also three other charts: the first shows the year-over-year changes for certain private market shelter indices in comparison to the CPI Shelter Index year-over-year. The second and third charts below show that if we adjust core CPI for our "market based" shelter index (second chart), year-over-year core CPI would be 1% (third chart).

Using various market-based private shelter trackers (ApartmentList, Realtor.com, Core Logic's Single Family Rental Index, and Case Shiller's Home Price Index), and taking an average, shelter prices on the ground are inflating at meaningfully lower rates than what CPI's Shelter Index is telling us. In recent months, it's actually demonstrating shelter prices are deflating (or decreasing).

Using the market based shelter index above in place of CPI Shelter, core CPI would be 1% in February year-over-year, about where it's been in previous months as well.

Lastly, let's try and assess inflationary breadth by looking at category level average and median changes. For those that normally read these posts, we do this by looking at "Category 4" and "Category 5" level details from the CPI's Table 2 from the Supplemental Table of Contents each month. Category 4 consists of 55 baskets of goods and services that comprise ~98% of the core CPI, while Category 5 consists of 101 baskets of goods and services that comprise about 78% of core CPI. By looking at the median and average changes in all these different categories, we can determine "inflationary breadth." If more of these categories are inflating at rates above 2% (which on a monthly basis would be something more than 0.2%), then it's easier to say that inflation in the economy is likely above the Fed's 2% target, whereas if categories on average are inflating at lower rates than that, we can say that it's the weight of the index that's causing the higher inflation readings.

The first two charts below show the monthly trends for goods and services in both Category 4 and Category 5. In February, Category 4 level goods and services increased 0.2% on both an average and median basis, while Category 5 goods increased 0.24% on an average basis and 0.3% on a median basis. Most of this is consistent with 2% inflation, especially once we adjust for the fact that this is CPI data, and the Fed tracks PCE, as we discussed above.

The last chart shows Category 4 and Category 5 averages on a 3 month average basis, and then annualizes them. This helps balance both recent trends while also minimizing for any individual monthly volatility. Annualizing these three month averages helps us see what category level inflation is (based on current trends) relative to the Fed's 2% inflation target. As the chart below shows, category level inflation remains well above what the index in general is saying inflation is (evidenced by the lines below the bars). Said again, this means that if we simply take the average and median monthly changes in a large basket of goods and services, we're at sub 2% inflation rates today.

All told, you'll continue to see headlines that inflation is stubborn and "stuck" above 2%, but it simply isn't true. Goods inflation has been higher than we'd like so far in 2024, but even despite that, we were able to print a CPI number that is consistent with 2% inflation, and that's despite the fact that CPI shelter remains meaningfully elevated relative to reality on the ground. We'd love for both core goods and core services ex. shelter to look more like they did for much of 2024 in coming months, but shelter remains the biggest problem with CPI, and not because of the data, but rather because of the methodology. Thankfully, the Fed is aware of this, but it can only do so much in spite of it because 1) the optics look contrary to this, and 2) ignoring one major category could lend itself to a slippery slope of picking and choosing which data to look at. Tariffs may complicate this picture in coming months, but for now, you should continue to be skeptical when you read about inflation in the press, because inflation in the United States is really in a much better place than most would have you believe.