Why US CPI Mattered—And Why it Didn’t

• Core CPI inflation was only seven one-hundredths lower than expected…

• ...but bond yields plunged, the dollar weakened, and equities rallied...

• ...probably mostly because they had gone too far in writing off Fed easing

• Core CPI suggests a second soft reading for the Fed’s preferred core PCE measure...

• ...but there are major reasons to doubt that inflation has been licked

• SA factors were high again, and distortions will continue in 2025

• US CPI / core CPI, m/m % SA, December:
• Actual: 0.4 / 0.2
• Scotia: 0.4 / 0.3
• Consensus: 0.4 / 0.3
• Prior: 0.3 / 0.3

CPI landed on the screws at 0.39% m/m SA (4.8% m/m SAAR). A slightly lower than expected core CPI reading of 0.23% m/m SA (seven one-hundredths lower than estimates) and a lighter, but not light 2.7% m/m SAAR (chart 1) motivated an outsized dovish market reaction. The reaction seems exaggerated and more to do with whippy markets, but reinforces the spillover effects from lower UK core CPI earlier this morning. 

US yields are down 9bps in 2s and 14bps in 10s on the day. The dollar is broadly weaker against all major crosses. Equities got a further boost on what was already a running head start assisted by strong beats across US bank earnings. Oh happy days indeed, or at least the singular version of that.

Why it Matters

Why it may matter is in terms of setting up a potentially second consecutive soft core PCE reading of underlying inflationary pressures when December’s tally arrives on January 31^st and following November’s core PCE reading of 0.1%. Back-to-back soft core PCE readings would be encouraging, but more evidence is needed in the context of plenty of forward-looking uncertainties and the current status of the US economy.

Core PCE often tracks softer than core CPI. There are many methodological differences between CPI and PCE. PCE captures more items with one glaring example being that it includes spending on goods and services by consumers (like CPI) but also on their behalf (like much the healthcare sector, and unlike CPI). The spending basket for PCE is updated each month to capture shifts in the composition of spending and hence inflation weights, whereas CPI adjusts weights once a year; hence, PCE does a better job at capturing substitution effects. There are other complicated differences but a key one is different spending weights.

Cue the estimate for PCE and core PCE using chart 5. Accounting for the different weights would have total PCE one-tenth below total CPI’s 0.4% reading. That probably also means core PCE will be one-tenth beneath core CPI’s 0.2% reading accounting only for the weighting differences.

Then add in what we learned from PPI that includes components included in PCE. They are shown in chart 6 and the outcome is that total PCE would get a 0.07 ppt lift from the relevant PPI components which rounds up to 0.1. Core PCE would get even closer to a 0.1 ppt lift from the PPI components because of adjusting weights to reflect the fact core PCE is a portion of total PCE and the chief driver of the PPI contribution to PCE was airfare which is included in core.

Did it Change Much for the Fed?

I wouldn’t say that the outlook for the FOMC’s actions spun on a dime on this one report and think this has more to do with markets perhaps righting themselves after going too far in writing off Fed easing for the year. Markets still have the FOMC parked on hold this month, added only 1bp to expected easing at the March meeting but that is still a long shot with only at most a one-in-four chance of a 25bps cut priced.

What instead drove the overall yield curve response was that markets went from one cut being priced for the year as a whole before CPI to nearly two cuts priced. The Scotiabank Economics’ forecast is for two Fed cuts this year, but there is enormous uncertainty around this call. There is substantial forecast revision risk at the start of every year, but especially the start of one like this given high uncertainty around the policy leanings of the incoming administration, among other factors.

You’ve been here before. In fact, plenty of times by now, as markets bounced all over the map while struggling to price a steady narrative.

For another point, we’re transitioning toward the start of the year when the recency bias in post-pandemic seasonal factors tends to overstate core CPI and PCE inflation readings. That did indeed work out as expected in December (chart 2), but it was offset by very slightly weaker than expected seasonally unadjusted core CPI. The next page includes charts 7–18 that show how SA factors in the pandemic era are higher than usual from about February through Spring, then ebb, and then return with a vengeance to close out the year. The recency bias in how seasonal adjustment factors are calculated means inflation can be overstated, then understated, then overstated again.

Furthermore, the bigger picture is unchanged. The US economy remains in excess aggregate demand and is performing quite well which should dampen some enthusiasm toward having licked inflation.

Fourth, there are plenty of upside risks. Cue Trump and on to next week’s developments.

The Details

Charts 3 and 4 shows that both core services (ex housing and energy services) and core goods inflation ebbed last month. There were multiple drivers shown in charts 19–30. Housing inflation was higher than the prior month, but this was offset by weakness across multiple other price signals. 

Also see charts 31–34 that break down the basket in y/y and m/m raw and weighted contribution terms. Finally, see the final page of this report that provides greater detail, more measures, and micro charts.