September 2026 CPI Report Results: Headline 0.6% MoM, 3.7% YoY

Headline CPI MoM

+0.6%, seasonally adjusted, after +0.4% in August. Gasoline rose 9.1% and accounted for over half of the monthly all-items increase; the energy index rose 5.2%.

Headline CPI YoY

+3.7% over the 12 months ending September, up from 3.4% in August — the annual headline rate re-accelerated. Energy is +20.4% year over year.

Core CPI MoM

+0.3%, seasonally adjusted, matching August. Shelter rose 0.4%; food was unchanged. Core excludes the energy surge that drove the headline.

Core CPI YoY

+2.6%, up from 2.4% in August. This is the series the core CPI-U markets resolve on; it printed above both pre-print co-favorites.

Headline YoY: re-acceleration priced

The annual market puts 86¢ on 3.6–3.7% against 3.4% in August — the board leans toward the year-over-year rate ticking up, with 3.5% at 7.2¢ and 3.8% at 7.4¢ as the next buckets.

Headline MoM: a hot month priced

The monthly market centers on +0.6% (48¢) and +0.5% (37¢) — together 85¢, above August’s +0.4% print.

Core: steady is the favorite

Core MoM favors 0.2% (50¢) over 0.3% (34¢) — a cooler month than August’s 0.3% print. Core YoY splits 2.4% (39¢) vs 2.5% (38¢) — the same 2.4% that won in August, priced at a coin-flip.

A fun sideshow

A separate market asks the average price of a dozen eggs in September ($31,585 volume so far). It resolves on a different series than CPI — a reminder to read every rule set before treating neighboring markets as one story.

Headline CPI MoM

+0.4%, seasonally adjusted, after +0.1% in July. Gasoline rose 3.9% and accounted for over one third of the monthly all-items increase; energy rose 2.1%.

Headline CPI YoY

+3.4%, before seasonal adjustment — unchanged from July. Energy is up 16.3% over the 12 months ending August; food is up 2.7%.

Core CPI MoM

+0.3%, seasonally adjusted, after +0.2% in July. This is the series the Polymarket month-over-month market resolves on; it came in one tenth above the 0.2% favorite.

Core CPI YoY

+2.4%, before seasonal adjustment, down from +2.5% in July — matching the bucket the year-over-year market had priced as its favorite.

The favorite that lost

The month-over-month market had priced 0.2% as the favorite (55.5¢ on September 8) — and 0.2% was also the most-traded bucket at $20,054 in lifetime volume. The printed 0.3% had been the 35¢ second favorite. A favorite losing is a normal market outcome, not a malfunction: prices are odds, not guarantees.

The YoY favorite won

The year-over-year market’s 2.4% favorite (46¢ on September 8, $17,006 bucket volume) matched the printed value after core inflation cooled from 2.5% to 2.4% on an annual basis — the gradual-disinflation path the board had priced most likely.

September-reference markets: settled in trading

All four September-reference markets settled in trading to the printed buckets on release day; formal resolution proposals were pending at the October 14 snapshot. Their pre-print boards and the settlement table are above; the October-reference events (settling December 10) are a separate contract window.

Headline CPI MoM

+0.1%, seasonally adjusted. Up from −0.4% in June. Shelter accounted for roughly two-thirds of the monthly increase; energy declined 1.5%.

Headline CPI YoY

+3.4%, before seasonal adjustment. Down from +3.5% in June. Energy rose 14.7% over the 12 months ending July, while food increased 3.0%.

Core CPI MoM

+0.2%, seasonally adjusted. Up from 0.0% in June. Medical care, airline fares, communication, education, and recreation increased; motor vehicle insurance decreased.

Core CPI YoY

+2.5%, before seasonal adjustment. Down from +2.6% in June. This is the same series family named by the reviewed year-over-year prediction market.

Core CPI MoM — 0.2% bucket wins

The 0.2% outcome bucket attracted the highest individual market volume at $56,808, well ahead of 0.3% ($39,842) and 0.1% ($38,444). Total event volume reached $173,068 with $298,764 in liquidity across all seven buckets.

View the resolved month-over-month event →

Core CPI YoY — 2.5% bucket wins

The 2.5% outcome bucket won with $19,032 in individual volume. Total event volume reached $139,959 with $331,470 in liquidity across all ten buckets. The 2.4% bucket ($45,352) was the most-traded non-winning outcome, followed by 2.7% ($20,828).

View the resolved year-over-year event →

A hotter core print than priced

Core CPI rose 0.3% month over month — one tenth above both July’s reading and the market’s 0.2% favorite. The annual core rate still edged down, from 2.5% to 2.4%, but the monthly acceleration matters for a Fed weighing a hike.

The Fed board repriced hard

At the September 11, 14:01 UTC snapshot — roughly 90 minutes after the print — the September Fed decision event showed 80.5¢ on a 25-basis-point hike versus 18.5¢ on no change, with $123.0 million in cumulative volume and $10.8 million traded in 24 hours. On September 8 the same board was a near coin-flip (52.5¢ hike / 46.5¢ hold).

Energy pushed the headline

Gasoline (+3.9% MoM, +27.4% YoY) and energy (+2.1% MoM, +16.3% YoY) drove the 0.4% headline print. Energy is excluded from core CPI, which is why the headline and core monthly numbers diverged this month.

What settles it

The September 15–16 FOMC decision arrived five days after the August print. The Fed event resolves on the actual policy move, not on CPI itself — see the September FOMC guide for the record.

October 27–28: the September print hit the Fed board

The October 14 report was the last scheduled CPI release before the October 27–28 FOMC meeting. At the October 14, ~14:00 UTC snapshot — about 90 minutes after the print — the Fed October event traded no-change at 59.5¢ versus 39.5¢ on a 25-basis-point hike ($38,327,996 cumulative volume). The headline re-acceleration (3.4% → 3.7% YoY) unwound most of the hold lead built after the September jobs report: the same board was 83/17 on October 12 and 82.5/16.5 on October 4.

Headline CPI, month over month

−0.4%, seasonally adjusted. BLS said the all-items decline was the largest one-month decrease since April 2020, with a 5.7% fall in the energy index the largest contributor.

Headline CPI, 12 months

+3.5%, before seasonal adjustment. This was the all-items change over the 12 months ending June.

Core CPI, month over month

0.0%, seasonally adjusted. "Core" here means all items less food and energy. This was the same series family named by the July month-over-month market.

Core CPI, 12 months

+2.6%, before seasonal adjustment. Down from +2.5% in July, showing the annual core rate fell over the month.

One-month change

The month-over-month event used the seasonally adjusted change in core CPI-U from June to July. It asked how much the index changed over one month after the seasonal adjustment specified in the rules. Answer: 0.2%.

Twelve-month change

The year-over-year event used the change in core CPI-U over the 12 months ending July before seasonal adjustment. It did not ask for the one-month figure annualized. Answer: 2.5%.

Same release, different cells

Both events named the August 12 BLS report, but they resolved from different values in that report. A correct answer to one contract did not determine the other.

One-decimal precision

Both rule sets say BLS reports the relevant core measure to one decimal point and that this is the precision used for resolution. More precise private estimates do not replace the published value.

1. CPI-U is specified

The reviewed events use the Consumer Price Index for All Urban Consumers. CPI-W, chained CPI, PCE inflation, producer prices, and private inflation estimates are different series.

2. Core is not headline

Both reviewed events specify all items less food and energy. The headline all-items number can move differently and cannot substitute for core CPI at settlement.

3. MoM and YoY use different comparisons

The one-month event compares July with June on a seasonally adjusted basis. The 12-month event compares July with July a year earlier before seasonal adjustment.

4. July is the reference month

The data month is July even though the release date is August 12. Record both fields so an August-release question is not confused with August data due in September.

5. Precision is one decimal

The rule text explicitly ties resolution to BLS one-decimal reporting. A model or newswire carrying more decimals is not the named resolution source.

6. Edge buckets are inclusive

The first and last buckets use "or less" and "or more." Middle buckets are exact one-decimal values. Check the actual group title rather than inferring boundaries from a chart.

7. The official BLS release controls

Both events name the monthly BLS CPI news release and say resolution takes place when the data is released. Commentary, consensus surveys, and market prices are supporting context only.

8. Delay fallback is explicit

If BLS does not release the relevant figure on schedule, the rules allow the market to remain open until the next CPI release time; if the figure is still unavailable, they point to the most recent prior month with available data.

Verify the release clock

Start with the live BLS schedule, record the reference month, and account for Eastern Time. Do not rely on a recycled calendar graphic.

Name the exact series

Write down CPI population, headline or core, comparison period, seasonal treatment, precision, and official source before opening an order book.

Read every outcome label

List the low edge, exact middle buckets, and high edge. Confirm inclusive wording and whether the market is one neg-risk group or separate binary books.

Timestamp market evidence

Capture bid, ask, spread, depth, volume, liquidity, and time together. Use the prediction-market odds guide before treating cents as a frictionless probability.

Map nearby catalysts separately

The 2026 jobs report calendar and September FOMC guide are useful context, but neither determines the CPI value or resolves these contracts. The October 27–28 Fed meeting guide extends the same calendar beyond September.

Keep the watchlist educational

The biggest-movers tool can flag repricing after the release. A move is a prompt to recheck facts, rules, and liquidity—not a recommendation.