US Core Inflation Would Make Upsizing Dicey

• Core CPI inflation is trending up with rising breadth…

• ...driven by core services and some tariff evidence

• Fed’s preferred core PCE is tracking on the fence between 0.2–0.3

•  Upsizing next week amid political pressure and data quality issues would be dicey

• US CPI / core CPI, m/m %, SA, August:
• Actual: 0.380 / 0.346
• Scotia: 0.3 / 0.3
• Consensus: 0.3 / 0.3
• Prior: 0.2 / 0.3

US core CPI inflation has accelerated over the past three months. This morning’s 0.346% m/m SA print was a smidge away from rounding up to a 0.4% core reading. Core PCE is on track for a 0.2–0.3% reading on September 26th after the Fed decision next Wednesday but they’ll have a good sense of the reading now. Such a core PCE reading would follow rounded-up 0.3% prints in the prior two months absent revisions.

Treasuries rallied somewhat on the back of the numbers and likely as a relief trade in that a higher than consensus number was just barely averted.

Next week’s FOMC meeting remains overpriced in my opinion. The first step toward easing is likely to be a careful -25bps cut, yet fed funds futures continue to lean toward a touch more. With financial conditions buoyant, GDP growth still tracking firmly as a correlated guide to dual mandate variables, and core inflation accelerating, it would be dicey to upside out of the starting gates amid data quality issues and political pressure.

DETAILS—BROADENING PRESSURES

The last three months of core CPI m/m SAAR inflation readings have been 2.8% in June, 3.9% in July and 4.2% in August (chart 1). That’s the warmest three-month moving average (3.65%) since January. What this tells us is that some combination of volatile relief and distorted seasonality created an optical illusion that inflation was ebbing in months like March and May. It is not.

Further, the breadth of price pressures continues to be rising again (chart 2). That’s a disturbing sign that price pressures are once again fanning out with half the CPI basket up by over 3% y/y and over one-third of the basket over 4%.

Prices for core goods ("commodities less food and energy") were up by 0.3% m/m SA which continues to be a mild acceleration (chart 3). This is the category where tariffs on goods would be most likely to show up which I’ll come back to.

Core services inflation (ie: ex-shelter and energy services) remained warm a 0.3% m/m SA after the prior 0.5% jump (chart 4) It’s running at 4% m/m SAAR. Who needs tariffs with core services inflation like that. This should be of greater concern to the Fed if it is inclined to look through tariff effects. Core service prices are more closely influenced by domestic economy considerations and are exhibiting significant stickiness.

Back to core goods prices. Chart 5 provides a weighted breakdown of the contributing drivers of this category. Some of these categories are obviously prone to potential tariff effects, like vehicles, swiss watches, imported tools/hardware etc. New vehicle prices were up 0.3% m/m SA but used vehicle prices were up a whopping 1% m/m SA. Used vehicle prices may also indirectly reflect tariff effects as used vehicles may be at a very early stage of exhibiting something akin to the pandemic when new vehicle prices were rising so folks went for used. 

In other words, I wouldn't shout if from the rooftops, but it would be equally inappropriate and arguably politicized to say there is zero evidence of tariff effects.

Chart 6 offers another way of looking at the category; core goods inflation is running toward the upper bound of experiences this decade.

The SA factor bias continues to ebb in 2024 and 2025 compared to earlier years (chart 7). I don’t judge this to have been a meaningful distortion this time as perhaps the BLS is reining in this effect. Chart 8 shows that the seasonally unadjusted increase in core prices was among the hotter readings compared to like months of August in history.

Implications for Core PCE Inflation

Converting to core PCE probably means a reading of 0.2–0.3% m/m SA. Taking account of weight differences in the two baskets (chart 9) plus the pertinent categories from yesterday’s PPI report (chart 10) puts the estimate close to the middle of that range. The wildcard is other methodological differences, different SA factors, and how the dynamic basket weights used in PCE versus CPI’s fixed annual weights translate into substitution effects.

Other Details

In addition to core services, other service categories were warm. Shelter inflation accelerated to 0.4% m/m SA. Rent of primary residence did likewise at 0.4% m/m SA which was the hottest since April (chart 11). OER was up 0.4%—the warmest since March (chart 12). Home insurance is also contributing (chart 13).

Charts 14–25 provide further breakdowns of components. Charts 26–27 break down the basket in m/m % changes by component and the weighted contributions to the m/m % change in total CPI. Charts 28–29 do likewise for y/y equivalents. See the accompanying table for more details.

We'll get the imputed share of the CPI basket using alternative methods just after 11amET.