The Big Idea
Getting to “greater confidence”
Stephen Stanley | February 2, 2024
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.
The January FOMC noted that “the Committee does not expect it will be appropriate to reduce the target range until it has gained greater confidence that inflation is moving sustainably toward 2 percent.” Dovish financial market participants surely were puzzled by that statement since the core PCE deflator has risen at a 1.9% annualized clip over the past six months and even slower over the past three months. Why is six months of data not enough? The composition of inflation in recent months suggests that the underlying trend is not nearly as benign as the core PCE deflator suggests.
Fed officials have been quite clear for at least a year that they are focused mainly on core inflation in tracking the underlying inflation trend. Food inflation has returned to a relatively normal pace of 2.7% in the CPI over the past 12 months, though the level of food prices is still quite elevated. In contrast, energy prices, as usual, have exhibited sharp swings. Over the 12 months of 2023, energy prices in the CPI fell by 2.0%, subtracting close to six tenths from the headline gauge. The year-over-year advance in the headline CPI was 3.3%, but the corresponding increase in the core CPI was 3.9%.
Within the core CPI, I have long singled out a group of five categories that have historically exhibited the most short-term volatility. I call them the Fearsome Five: new and used vehicle prices, apparel, airfares, and hotel rates. In total, they add up to about a 13.5% weight in the core CPI. The swings in these line items typically bear little connection to underlying inflation trends.
Starting in June, this group has posted substantial declines in prices. The Fearsome Five subtracted from the core CPI for six months in a row through November before posting a small positive contribution in December. That swing was a key reason why the core CPI rose by 0.3% in December. Over the past seven months, the Fearsome Five has cumulatively subtracted 0.3 percentage points from core CPI inflation, while the rest of the core has risen at a 3.5% annualized pace.
There are several other alternative ways to approach the CPI data. The Cleveland Fed calculates a Median CPI and a Trimmed-Mean CPI. For each, the first step is to rank all of the line items from highest inflation to lowest. Then, the Trimmed-Mean CPI lops off the top 8% and bottom 8% of the index—each line item counts for its weight in the index—while the Median CPI isolates the single line item that falls in the 50th percentile of the rankings.
The Trimmed-Mean CPI has been running close to the core CPI. Both are up 3.9% on a year-over-year basis. But the Median CPI has been increasing much faster. In fact, the Median CPI has risen by at least a full tenth faster than the core CPI in each of the past four months, and the year-over-year advance in the Median CPI is an astounding 5.1%, more than a full percentage point above the core CPI.
The Atlanta Fed offers an alternative approach called the Sticky-Price CPI. The intuition behind this index is that for goods and services whose prices change infrequently, sellers will only raise or lower prices when they see a persistent trend that justifies such a move. While goods and services with flexible prices—in the extreme, think gasoline pump prices—may gyrate up and down frequently and wildly, the moves in sticky price items offer a more accurate window into underlying trends. In any case, the Sticky-Price CPI is showing a faster trend than the core CPI, 4.6% over the past 12 months and 4.2% annualized over the past 3 months.
Core PCE versus Core CPI
The core PCE deflator is, of course, the gauge that Fed officials follow most closely. Historically, the core PCE deflator has tended to run below the core CPI, with the historical gap averaging just under half a percentage point. However, right now, the difference is much wider, a full percentage point over the past 12 months. While the core CPI is up by 3.9% over the past 12 months, the core PCE deflator has risen by only 2.9%, much closer to the Fed’s liking.
Moreover, the gap has been even wider in recent months, running at 135 bp on a 6-month annualized basis and 181 bp on a 3-month annualized basis. Pinpointing the drivers of this yawning discrepancy in the second half of last year may offer hints at how it will close. Will the core CPI converge downward toward the core PCE deflator or will the core PCE deflator move higher toward the core CPI?.
Some of the sources of the unusual gap are understandable and not necessarily fleeting. For example, shelter costs represent about 40% of the core CPI but have a weight of only about 15% in the core PCE deflator. A 6% rise in shelter costs, close to the most recent 12-month advance, is worth 240 bp for the core CPI but only about 90 bp for the core PCE deflator. This probably explains the largest portion of the unusually large difference between the two.
There are other sources as well. For example, the methodologies for calculating motor vehicle insurance rates differ. The CPI measures the cost of a constant coverage package to a household. Motor vehicle insurance rates have been skyrocketing in recent years, as the costs of repairs surge. The motor vehicle line item in the CPI is up 20% over the past 12 months, adding roughly three-quarters of a percentage point to the core CPI advance. In contrast, the BEA defines motor vehicle insurance prices based on the difference between premiums taken in by carriers and claims paid out. This may not seem intuitive, but they have to do it this way to square up the GDP calculations. By this method, motor vehicle insurance prices are still up a hefty 9% over the past year, but the expenditure weight is far smaller since only the net payments are being tallied. As a result, the contribution to the core PCE deflator is a mere 5 bp.
In sum, it is not obvious that the unusually large gap between the two gauges of core inflation is going to close any time soon. If anything, it seems that the core CPI could converge toward the core PCE deflator, but the disposition of this gap introduces yet another element of uncertainty for policymakers trying to gain confidence that the underlying trend is moving sustainably to 2%.
Underlying trends for core PCE
As with the core CPI, there are alternative calculations for the core PCE deflator that offer additional insight into the underlying state of inflation. The Cleveland Fed calculates a Median PCE gauge in the same way that it crafts the Median CPI. That measure is running at 3.8% on a year-over-year basis in December, nearly a full percentage point above the corresponding core PCE deflator reading of 2.9%. Over the past six months, the difference is even larger, 3.1% annualized for the Median PCE vs. 1.9% for the core PCE deflator.
Similarly, the Dallas Fed Trimmed-Mean PCE gauge is running at 3.3% over the past 12 months, four tenths of a percentage point above the core PCE deflator, and the six-month annualized increase is 2.6%, over a full percentage point higher than the corresponding core PCE deflator result.
In sum, the alternative gauges of underlying inflation offer a compelling case that the low core PCE readings in recent months should be viewed skeptically. All of these measures are running far above the corresponding core PCE gauge – and far above the Fed’s 2% target.
At first glance, it would seem that six months’ worth of core PCE inflation running slightly below the Fed’s 2% target would be ample evidence to support the beginning of rate cuts. However, a variety of alternative measures of underlying inflation show consistently that the moderation in inflation over the past year has been relatively narrowly based and dependent on some idiosyncratic and likely temporary forces. Every one of the alternative gauges that I have highlighted in this piece are tracking far higher than their more traditional counterparts.
I have been making the case for months that the aggregate inflation figures were overstating the degree of progress in the true underlying trend. While the FOMC and Chair Powell did not explicitly say so, I suspect that their reticence to declare victory on inflation reflects a recognition that things are not quite as good as they may appear. Historically, when core inflation measures diverge substantially from the corresponding trimmed-mean and median calculations, since it reflects by definition that the inflation trends are being driven by a narrow set of line items, the core indices tend to converge back to the alternative measures, not the other way around. This is what I expect to see in early 2024. Though the underlying pace of inflation is probably still decelerating, I expect the core inflation prints over the next several months to accelerate from the surprisingly low readings seen in the second half of last year. In this scenario, financial market participants may have to wait far longer than they currently believe before Chair Powell and the FOMC get to “greater confidence.”