Consumer Prices Come In Softer Than Expected; Tariff Passthrough Still Moderate; Service Inflation Still Elevated

Finally, some data!

“I sent for yesterday but here you come today!”
—Count Basie

It’s nine days late, thanks to the government shutdown, but the BLS released September’s Consumer Price Index this morning, giving market analysts, the Fed (which meets next week), businesses and the rest of us a much-needed-look at a key data point. Price growth came in a bit softer than expected, locking in a Fed rate cut next week and possibly setting them up for another cut in December. Tariffs are still in the price data, but passthrough remains subdued.

My main concerns at this point are a) tariff passthrough to come, and b) service price pressures.

The topline numbers, along with market expectations and last month’s rates, are in the table below. Some highlighted observations follow.

—Inflation rose 3% year-over-year for both headline and core, both a touch below expectations.

—Monthly rates also were both a tenth below expectations.

—Markets responded positively to the softer-than-expected print, with equities up and yields down. An October rate cut was already in the cards, but with this report, it’s as locked in as they get.

—Under the hood, there are some mixed signals. Gas was up 4.1% for the month, its highest rate since August of ‘23, but that’s a jumpy number. Apparel prices were up 0.5% in Aug and 0.7% in Sept, likely reflecting tariffs. But used cars fell 0.4%.

—Core services ex-housing (see more below) was up 0.4% (0.35%) in Sept and 3.2% yr/yr. This series remains elevated—pre-pandemic it ran closer to 2% than 3%—and continues to deter inflation’s path back down to the Fed’s 2% target.

I’ve been paying close attention to this services ex housing category, as I’ve seen pressures there in recent months that arguably are not directly linked to tariffs, as they fall on (imported) goods. My go-to technique when I’m trying to suss out not where just inflation is but whether its speeding up or slowing down (first and second derivatives, for you calculus fans) is to look at annualized growth rates at 12, 6, 3, and 1 month intervals. If the bars are tilting up as you go from left-to-right, as they kinda are here, it suggests inflation acceleration, as nearer-term intervals are posting higher rates.

The main reason the Trump administration called back the workers needed to release the September data is that it is needed to calculate the annual cost-of-living adjustment (COLA) for Social Security benefits. Today’s release returns a COLA of 2.8%, which will be applied to December 2025 benefits, payable in January 2026 (the COLA is calculated using Q3/Q3 price data from the CPI-W subindex). The average monthly benefit increase due to the COLA should be ~$56.

For all the mayhem, chaos, trade war, shutdown, violations of the rule-of-law, and figurative and now literal wrecking the foundations of democracy (you see the East Wing? Me neither…), it’s comforting to see that insurance against the economic impacts of aging and retirement, i.e., Social Security, is still intact and that benefits are on track to be updated for inflation.

Commentary:

Even as the report came in softer than expected, there’s no question that yearly CPI inflation has been accelerating, and that the core is stuck well above the Fed’s target. The moderation in tariff passthrough has been a boon, no question, but the figure below, in tandem with the clearly weakening job market, suggests our stagflation threat is ongoing.

Someone asked me why I print the expectations. In fact, they’re one of more important aspects of this and other such reports, especially now, as we’re trying to assess the price damage from tariffs. What the expectations tell you is how well we understand what’s going on with prices. Any month can surprise you a bit up or down, but as long as our expectations are in the ballpark, we can take comfort that inflation is moving as we thought it was, given our understanding of the forces that shape it. That’s particularly important for the Fed, as they need to set policy based on where they thing inflation and employment are going, and big surprises call their understanding into question.

Recently, the actual and expected reading on inflation have been close, rarely off by more than a tenth-of-a-percent. This is not unusual, in part because the inflation data are “autoregressive,” i.e., they have a memory such that inflation at time t, absent a big shock, tends to not drift too far from inflation at time t-1.

Here’s how most of the Fed board is thinking about inflation and their rate-cut path right now. This is from Philadelphia Fed president Anna Paulson, last week: “My base case is that tariffs will increase the price level, but they won’t leave a lasting imprint on inflation. And, given this base case, monetary policy should look through tariff effects on prices.”

This gives them the space to focus on the weakening labor market, which says “cut” vs. sticky inflation, which says “pause.” The challenge with this stance is that it requires the Trump admin to settle into a tariff regime, enabling the Fed to “look through” the one-time import tax. But he won’t do so—see his latest hissy fit this AM re Canada—which undermines the one-time aspect of the shift.

But there’s another problem, a bigger one from consumers perspective. It’s the “increase the price level” point made by Ms. Paulson. What consumers have been most upset about in recent years isn’t so much inflation as the price level, i.e., how much things cost. The tariffs, even with the quite moderate passthrough we’ve seen, push the wrong way on that.

While the Fed may look through price levels to price changes, the consumer stops at the price level, and she’s not happy about it.