August CPI Beat Came From Fuel, Not Tariffs

Markets spent months braced for the wrong thing. The consensus fear ahead of the August inflation report was tariff pass-through: import duties finally showing up as higher prices on core goods. When the beat landed, it came from a completely different corner of the economy.

The August 2025 Consumer Price Index release, published by the Bureau of Labor Statistics on 11 September 2025, showed core CPI rising 0.3% month-over-month against a consensus of 0.2%. Headline CPI came in at 0.4% month-over-month and 2.9% year-over-year. The upside driver was not goods at all. It was services tied to fuel costs, led by airline fares up 5.9% in a single month, alongside a shelter reading that re-accelerated to 0.4%. Core goods, the tariff-exposed part of the ledger, barely moved.

Here is what that split actually tells you, and why it matters more than the headline number. This piece gives you a working framework for reading the next release: which categories carry the signal, why the goods-versus-services divide changes the policy stakes, and how a diesel price feeds inflation that officially excludes energy.

What the August numbers actually showed

Start with the clean figures. According to the BLS, headline CPI rose 0.4% in August after 0.2% in July, landing at 2.9% year-over-year. Core CPI, which strips out food and energy, rose 0.3% on the month and 3.1% over the year. That core figure is the one that beat, exceeding the 0.2% the market expected.

A note on the numbers before going further. Some secondary commentary circulated softer annual figures for August (around 2.4% core and 3.4% headline). Those do not reconcile with the BLS release and are not used here. The primary statistical figures are 3.1% core and 2.9% headline.

The interesting part is not the beat itself. It is where the beat came from.

The composition behind the print

Rank the contributors by weight and the story writes itself. Shelter did the heavy lifting, rising 0.4% on the month, and the BLS identified it as the single largest contributor to the all-items increase. That matters because shelter represents more than 40% of core CPI, so even a modest monthly move drags the whole index with it.

Airline fares were the standout mover in percentage terms, jumping 5.9% in August. The energy index rose 0.7% on the month but sat almost flat over the year at 0.2%, which tells you the energy story here is recent and sharp rather than a slow annual grind.

Then there is the category the market was watching most closely. Core goods rose just 0.1% on the month and 0.7% over the year, with no visible tariff signal.

The composition is the read. This was a services beat with two very different persistence profiles underneath it, and separating them is what turns a headline reaction into a useful one.

How diesel prices end up inside core CPI

Airline fares jumping 5.9% in a month is a striking number on its own. The more useful question is why a fuel-driven cost is showing up in a measure that explicitly excludes energy at the commodity level.

The answer is transmission. Fuel does not enter core CPI as a barrel of jet fuel or a gallon of diesel. It enters as a higher service price, once businesses that burn that fuel pass the cost on to customers.

Take the airline fares channel step by step, using BLS methodology:

1. The CPI airline fares index tracks actual ticket prices paid, including the base fare, taxes, carrier-imposed fees, and fuel surcharges.
2. When jet fuel prices climb, carriers respond by lifting base fares, adding or raising fuel surcharges, or trimming capacity to support higher pricing.
3. Because the index captures real transaction prices, those increases flow into measured core CPI with a short lag once the new fares are widely in effect.

Ground transportation works the same way, through a different set of contracts:

1. Diesel prices rise at the pump and across freight networks.
2. Trucking, delivery, and ride-hailing operators pass that through via fuel surcharges in freight contracts and higher per-mile charges.
3. Those higher fees land in the CPI transportation services category as more expensive services, not as an energy commodity.

That is the mechanism that resolves the apparent contradiction.

Energy is excluded from core CPI at the commodity level, but when energy raises the cost of delivering a service, that higher service price is captured in core CPI.

Energy costs embedded in core CPI through service prices create a systematic measurement gap: the Dallas Fed estimates the 2026 Iran conflict raised US headline PCE by approximately 0.6 percentage points through this channel alone, with war-driven fuel costs appearing as transportation and airfare price increases rather than as the energy index entry they economically represent.

A word on the diesel figures, because the record-high claim circulating in some coverage does not hold up. According to Energy Information Administration weekly data, the verified all-time national average record for on-highway diesel is $5.81 per gallon, set the week of 20 June 2022. A figure of $6.05 per gallon appeared in some August 2025 commentary, but it is not confirmed by EIA data and carries no named source. Recent national averages have run closer to $5.60 per gallon.

EIA weekly diesel price data tracks national and regional on-highway averages with enough granularity to verify whether spot moves in fuel have peaked, plateaued, or begun reversing, making it the primary reference for calibrating the fuel-to-services pass-through thesis going into the September release.

The practical implication runs forward. Energy price moves can push core inflation higher even when goods tariff pass-through is completely absent. That makes fuel futures a leading indicator for future core services readings, and it means the August services pressure is unlikely to be a one-month anomaly if diesel and jet fuel stay elevated through September and October.

Services inflation versus goods inflation, and why the distinction matters for the Fed

Two types of inflation can produce the same number on a headline index while pointing at completely different policy futures. That is the real tension inside the August print, and it explains why bond markets and Fed officials react differently depending on which side of the ledger is driving.

The headline-versus-core gap is the primary diagnostic tool across recent inflation reports: in May 2026, a 1.3 percentage-point spread between the two measures confirmed that energy was doing the heavy lifting rather than broad domestic demand, which is the same analytical frame that makes August 2025’s composition readable rather than ambiguous.

Fuel-driven services inflation is tied to spot and futures energy prices. Those are volatile and can reverse quickly, which makes this kind of pressure cyclical and potentially transient if fuel retreats. Tariff-driven goods inflation is a different animal: a tariff is a policy action that stays in place until it is changed, so its price effect tends to be more durable.

The Federal Reserve uses core CPI as a working measure precisely because direct energy prices are too volatile to steer policy by. The wrinkle is that fuel-driven service costs still land inside core CPI, and they can weigh on Fed deliberations when they persist rather than fade.

Read against that framework, August is a mixed signal. Core running at 3.1% year-over-year sits well above the 2% target, which keeps it policy-relevant. But the driver is services tied to an energy spike, not a structurally sticky source like wages or rents resetting permanently higher.

Three features of the August data point to services-led rather than goods-led pressure:

• Core goods were quiet at 0.1% month-over-month, a clean negative on tariff pass-through.
• Airline fares spiked 5.9%, the visible fuel-to-services channel at work.
• Shelter cooled to 3.6% year-over-year, its lowest reading since October 2021, even as it re-accelerated on the month.

The St. Louis Fed attributed approximately 0.5 percentage points of annualised headline PCE inflation over June-August 2025 to tariffs, which puts the August 2025 core goods reading of 0.1% month-over-month in useful context: tariff pass-through has been measurable in aggregate data without being visible in a single month’s category breakdown.

What this tells you is that the core beat is not automatically evidence of re-accelerating inflation. Whether it sticks depends almost entirely on what energy prices do over the next two months, which is why watching energy markets now matters as much as watching labour data.

Shelter re-acceleration and what it adds to the picture

Shelter is the quiet co-driver that complicates any tidy “this was just a fuel blip” reading. Even as the long-awaited shelter disinflation continues on an annual basis, a single strong month was enough to keep core CPI above the comfort zone.

The weight is the reason.

Shelter constitutes more than 40% of core CPI, so even a modest monthly move has an outsized effect on the core reading.

In August, shelter rose 0.4% on the month, and the BLS flagged it as the largest single contributor to the all-items increase. That is a genuine re-acceleration from softer prior readings, and it is a large part of why the core print landed above consensus rather than at it.

The annual figure pulls in the other direction, and that tension is the whole point. At 3.6% year-over-year, shelter inflation is the lowest since October 2021, per NAHB/EyeOnHousing data, which confirms the multi-year disinflation trend is still intact. Monthly volatility around that trend is real, and it cannot be waved away in any single report.

For your read on the next release, the shelter story breaks into three parts:

• The current monthly reading of 0.4%, well above the pace consistent with target inflation.
• The annual trend direction, still falling and encouraging.
• The threshold that matters: a stubbornly elevated monthly reading keeps core CPI above the Fed’s comfort zone regardless of what fuel does.

That last point is the actionable one. Even if airfare spikes reverse in September, shelter holding at an elevated monthly pace is sufficient on its own to keep core elevated. Shelter is also far more predictable than energy, so tracking rent and owners’ equivalent rent trends gives you a structural edge in anticipating where core CPI heads once the fuel noise clears.

What the August data does and does not change about the inflation outlook

Two forces were at work in August, and calibrating between them matters more than reaching for a verdict. Fuel-driven services inflation is potentially reversible if energy prices retreat. Shelter disinflation is directionally encouraging on the year but is throwing off enough monthly volatility to keep core elevated.

Be precise about what the print does not establish. It does not confirm tariff pass-through, because core goods were quiet at 0.1% month-over-month. It does not signal structural re-acceleration, because shelter is still falling year-over-year at 3.6%. And it does not, on its own, change the Fed’s direction; at most it affects the pace.

Core running at 3.1% against the 2% target keeps the pressure real. Empower’s framing of inflation as “close to 3% as the Federal Reserve considers its next move” captures the calibrated reality: this is a trend-context data point, not a decisive single-month pivot.

That leaves three variables to watch when the September release arrives:

1. Airline fares: a reversal from the 5.9% August spike would be the clearest sign August was fuel-driven noise.
2. Shelter: a stubbornly elevated monthly reading would keep core elevated independent of fuel.
3. Core goods: any material move above 0.1% would be the first real sign of tariff pass-through.

Read August as a complexity report, not a verdict. The next two prints will tell you far more about whether fuel-driven services pressure fades or turns into the opening act of a new services-inflation cycle. Rather than repositioning around one number, watch which of the three variables moves first.

Reading the next CPI release with the right lens

The lasting value of the August data is that it hands you a baseline decomposition: services-led, fuel-driven, goods quiet. September’s release becomes something you read comparatively against that baseline rather than in isolation.

The central question September must answer is simple: was this a one-month fuel spike, or the start of a sustained energy-to-services pass-through cycle?

Turn that into a working checklist for when the September print drops:

1. Did airline fares reverse from the 5.9% August spike? A pullback supports the transitory read on the fuel channel.
2. Did shelter remain elevated on the month? Continuation keeps core pressure alive regardless of energy.
3. Did core goods move materially above 0.1%? That is the threshold before tariff pass-through becomes a credible competing narrative.

Enter the September release with those three thresholds pre-defined and the report stops being an anxiety event. It becomes an analytical exercise where you already know what each answer signals. That is the difference between reacting to a headline and reading the data.

For investors wanting to see the fuel-to-services channel run in reverse, our deep-dive into June’s core CPI miss traces how falling gasoline prices drove the sharpest monthly headline decline since April 2020, providing a concrete before-and-after comparison for how quickly fuel-driven core pressure can unwind.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results, and forward-looking interpretations are subject to market conditions and various risk factors.