August Inflation Is an Oil Shock. The Fed Should Hold.

Headline CPI and producer prices both firmed in August. Strip out energy and there is little evidence of a broad reacceleration — which is exactly the distinction that should shape the Fed's decision next week.

Key Takeaways

• Headline firmed, and it was energy. CPI rose 0.4% in August, up from 0.1% in July; gasoline alone was more than a third of the monthly increase. Core CPI eased to 2.4% year-over-year.
• The pipeline tells the same story. Producer prices rose about 0.4% in August and 5.4% over the year, but the clear accelerants were energy and freight. Final-demand energy rose 4.2% in the month, while transportation and warehousing services rose 2.3%. The PPI measure excluding food and energy slowed month-over-month.
• The Fed’s preferred gauge should stay contained. My CPI-and-PPI bridge model nowcasts core PCE at +0.22% month-over-month for August — enough, on the currently published PCE history, to hold the annual rate near 3.3%, with three-month momentum running cooler, around 2.5%.
• This isn’t demand to lean against. Inflation is still too high. But with a narrow labor market, moderating wage growth, and anchored expectations, the Fed should hold.

What happened

Two inflation reports landed this week, and read together they confirm one thing. Consumer prices firmed. Producer prices firmed. And in both, the new inflation impulse came primarily from energy and the freight that energy moves — not from a broad revival in demand.

Start with the consumer side. The all-items CPI rose 0.4% in August, up from 0.1% in July, leaving the annual rate at 3.4%.

The single biggest driver is unambiguous: by the Bureau of Labor Statistics’ own accounting, gasoline was more than a third of the entire monthly increase. Gasoline rose 3.9% on the month and is up 27.4% over the year; the broader energy index rose 2.1% in August and 16.3% over twelve months.

Oil back above $100 on Middle East supply risk is flowing straight to the pump.

Now strip energy out. Core CPI — all items less food and energy — eased to 2.4% year-over-year, down from 2.5%. The month did carry a 0.3% core print, but some of the firmness sits in volatile corners: airline fares rose 2.7%, communication 2.3%, and lodging away from home 2.4% after falling 2.8% the month before.

Owners’ equivalent rent, the heart of the shelter index and one of the stickiest pieces of the basket, rose just 0.2%, and shelter’s annual rate cooled to 3.0% from 3.2%.

This is not a benign inflation report. But neither is it evidence of a broad re-acceleration. Shelter continues to cool, while much of the new pressure is showing up in energy, travel, and other volatile components.

The producer report rhymes. Final-demand PPI rose about 0.4% in August and is up 5.4% over the year — a hot-looking number until you open it.

Final-demand energy prices jumped 4.2% in the month. Transportation and warehousing services rose 2.3%. Diesel alone surged more than 24%.

Those categories were the clear accelerants.

The PPI measure excluding food and energy slowed to roughly a 0.2% monthly pace, while the broader measure excluding food, energy, and trade services rose 0.3%, down from 0.4% in July. That measure remains elevated at 4.7% over the year.

In other words: producer inflation is still too high, but the August acceleration was concentrated exactly where a supply shock would put it.

Why does the freight line matter? Because energy is a major input into transportation, and transportation costs can work their way through the prices of goods moving through the economy.

That is the channel to watch: energy into freight and, potentially, into broader consumer prices. It is a cost-push risk, not evidence that households are suddenly spending the economy hot.

What it means for households

For households, the firm-up is real and it is regressive. Gasoline is the most visible price in the economy, and it is rising fastest.

Layer that on a consumer who is already stretched. Average hourly earnings are up about 3.1% over the year and steady — not accelerating. BLS estimates that real average hourly earnings fell 0.3% over the year and 0.1% in August.

The household is on a treadmill, running to stay in place.

The squeeze is also coming through borrowing costs. The sharp rise in long-term Treasury yields has already tightened financial conditions — mortgages, auto loans, and credit are all more expensive — without the Fed touching its policy rate. Add the oil tax on top, and the consumer is being hit from several directions at once.

The cushion that would normally absorb this is thin, because the labor market’s strength is narrow. The headline unemployment rate looks fine, but the long-term unemployed — those out of work 27 weeks or more — are now about 1.1% of the labor force, roughly 50% above the level implied by a healthy expansion at today’s unemployment rate, and about 27% of everyone unemployed.

The problem isn’t a wave of layoffs. It is that once workers are out, they struggle to get rehired. A labor market like that does not have much room to take a policy mistake.

What it means for the Fed decision next week

The Fed doesn’t target CPI or PPI. It targets PCE — and the August PCE report doesn’t publish until September 30, after next week’s meeting.

So this week’s consumer and producer prices are the committee’s real-time read on where its preferred gauge is heading. That read is the useful output of this exercise.

I estimate a simple bridge: regress monthly core PCE on the same-month core CPI and core-services PPI inputs that release before it, and nowcast the pending month. In a walk-forward pseudo-out-of-sample exercise, the full CPI-plus-PPI model beats a CPI-only version and a random walk. Its RMSE is about 0.080, versus 0.085 for the CPI-only model and 0.128 for the random walk. For August it nowcasts core PCE at +0.22% month-over-month, with an 80% band of roughly [+0.12, +0.32].

Two implications follow.

First, on the year-over-year rate: on the currently published PCE history, the August 2025 month now rolling out of the calculation was itself about +0.22%. So a +0.22% August 2026 print would leave core PCE holding near 3.3% rather than climbing.

Three-month annualized momentum, around 2.5%, is running below the annual figure. The underlying trend is easing.

Second, headline PCE will run hotter, lifted by the same gasoline surge that drove CPI.

There is one important caveat. The September 30 PCE release coincides with the BEA’s annual update, so previously published history can be revised. That means the realized year-over-year rate may move even if the August monthly print lands close to the nowcast.

This is where the composition of inflation matters for policy.

The San Francisco Fed’s decomposition attributes only about 40% of current inflation to demand-driven categories; the rest is supply-driven or ambiguous.

Monetary policy works most directly on demand. When the marginal impulse is a barrel of oil, a rate hike won’t unclog a supply chain or unwind a tariff. It tightens the screws on demand that is already softening.

The committee has told us how it will think about this. The July FOMC minutes laid out the competing risks: inflation could remain elevated because of Middle East supply disruptions, tariffs, or AI-related demand, while the labor market remained an important part of the policy calculus.

That labor-market assumption is becoming less comfortable.

Long-term unemployment is elevated relative to a healthy expansion. Employment gains remain concentrated. Wage growth continues to slip. The labor market is not collapsing, but neither is it providing the kind of buffer that makes another tightening move costless.

Another reassuring piece for the Fed is that expectations remain contained.

Household expectations are elevated but stable: roughly 3.2% at three years and 3.0% at five years. Market-based measures are calmer still, with five-year inflation compensation around 2.5% and longer-run forward compensation near 2.3%.

Consumers see a bumpy inflation path. Markets are not pricing in a spiral.

Put it together, and the case for patience is stronger than the hot headline suggests.

Inflation is still too high. The Fed should not look through every supply shock forever. If higher energy and freight costs begin spreading into broad core inflation, wage growth turns back up, or expectations start moving materially higher, the calculus changes. But that is not what August shows.

For now, the inflation problem is getting hotter where monetary policy has the least leverage, while labor and financial conditions are already tightening where it has the most.

Hold the rate. Watch the pass-through. Inflicting more pain on this economy won't cure the oil shock.

Sources. U.S. Bureau of Labor Statistics, Consumer Price Index — August 2026, Producer Price Index — August 2026, and Real Earnings — August 2026. U.S. Bureau of Economic Analysis, Personal Income and Outlays, July 2026; next release covering August scheduled for September 30, 2026. Federal Reserve Bank of San Francisco, Supply- and Demand-Driven PCE Inflation decomposition. July 2026 FOMC minutes. New York Fed Survey of Consumer Expectations and market-based inflation-compensation measures. Core PCE nowcast and figures are the author’s estimates using BLS and BEA series via FRED; nowcast bands reflect pseudo-out-of-sample model error and are not official forecasts.