September CPI October 14: The Print That Decides the Fed's December Vote

October 13, 2026 · 8 min read · ChartWise Team

September CPI lands Wednesday at 8:30 AM ET (6:00 PM IST), and it is the last inflation reading the Fed gets before its final rate decision of the year. The consensus expects headline inflation at 0.5% month over month and 3.6-3.7% year over year, with core at 0.2%. Even that expected path lifts the annual rate from 3.4% toward 3.7%, because energy prices keep climbing. The November CPI does not publish until December 10, one day after the December 9 decision, so this print carries the year's last real vote on rates all by itself.

Where the Market Stands Going In

Gold spent last week doing nothing wrong and still lost ground. Futures closed Friday, October 9, at $4,216 and sat near $4,188 early Tuesday, with spot trading around $4,165 in Asia. That is a $50 drift lower into a major data print, on top of a September that took nearly 6% off the metal and left it roughly $1,400 below January's record high of $5,608.

The rates tape explains most of it. The 10-year Treasury yield closed Monday at 5.24%, just under the 5.31% it touched October 5, the highest daily reading since May 2002. The dollar index sits at 102.5, up from about 99 in mid-September, which is a 3.5% dollar rally in a month. EUR/USD has given back the whole late-summer bounce and trades around 1.12 after printing 1.16 in early September. The Dow closed Monday at 51,850 and the Nasdaq at 27,246, both short of their September highs but nowhere near distressed.

Which means the market walks into this CPI already paying for a hawkish outcome. Yields near multi-decade highs, a strong dollar, and 70% December odds all say the hot scenario is partly pre-paid. That was exactly the setup two weeks ago before core PCE, and the same asymmetry applies: a hot print confirms pricing and extends rather than gaps, while a cool print contradicts a market that just spent a month pricing the opposite. Contradicted markets move violently.

The September CPI Forecast

CPI measures the change in prices urban consumers pay for a basket of goods and services. Core CPI strips out food and energy because they swing around for reasons (weather, geopolitics, OPEC meetings) that tell the Fed nothing about domestic demand. The Fed targets PCE inflation, but CPI moves markets first because it lands first, and the September PPI follows a day later this month instead of the usual day before.

Put the annualization math on the core cluster, because that is where the trade lives. Deutsche Bank's 0.16% street-low annualizes to 1.9%. Natixis at 0.19% is 2.3%. The 0.21% median is 2.5%. August's actual 0.29% was 3.5%. The entire forecast column sits below the prior print, which is unusual and important: consensus is not just expecting cooling, it is expecting the August print to have been a one-off. Natixis says exactly that, calling the August beat a rise in components that should not repeat. If August does repeat, the move will be large, because almost nobody has it in their number.

Why This Print Decides December

Four weeks ago the Fed raised rates a quarter point to 3.75%-4.00%, its first hike since 2023, on a unanimous 12-0 vote. The September minutes, published October 7, showed most participants judged that "one more increase by year-end" would be appropriate, and 16 of the 18 officials who filed projections put the year-end rate at least a quarter point above the current midpoint. The committee is not hiding where it wants to go. It is waiting for permission from the data.

The calendar then does something cruel. November CPI publishes December 10, the day after the December 9 decision, so the Fed will vote on rates without seeing it. That leaves this CPI and Thursday's PPI as the committee's final clean inflation reads of the year. September jobs, released October 2, already removed most of the argument for October: payrolls rose just 29,000 against forecasts near 90,000, the unemployment rate ticked up to 4.2%, and the two prior months were revised down a combined 60,000. Wage growth slowed to 3.0% year over year, the weakest since May 2021. Jefferies called the report the nail in the coffin for an October hike, and FedWatch pricing agrees, with a hold on October 28 above 80%.

That leaves a squeeze. The labor side of the mandate is cooling on schedule, while the price side accelerates on energy nobody in Washington controls. The Fed gets one more vote this year, the minutes say most members want to use it, and the deciding evidence arrives Wednesday at 8:30. A core print of 0.2% likely turns December into a genuine coin flip and reprices it toward 50%. A core print of 0.3% likely pushes December above 85% and puts 25 basis points on with nobody surprised.

The Blueprint: How the Last Hot CPI Actually Traded

Size Wednesday by what gold actually did on its last three repricing days, because the pattern contradicts what most traders expect from a CPI.

August 12, in-line July CPI: gold ran about $55 an ounce to a $4,438 high on a print that surprised nobody, and September hike odds fell from 46% to 40%. Even agreement moves this market when positioning is nervous. September 11, the hot August CPI: core came in at 0.3% against 0.2% expected and the annual rate beat at 3.4%. Gold futures closed within $2 of the prior day. On the tape, the hottest CPI of the cycle did nothing. But September hike odds jumped from about 56% to 84% by that afternoon, per Goldman Sachs' own trading desk, and by September 15 gold had slid to $4,333, roughly $75 below the print-day close. The Fed hiked on the 16th, and the decision day itself produced a bounce, with gold closing near $4,388. Across September, gold lost nearly 6% even though the big print day was flat.

September 28-29 complete the picture: no data release at all, just yields and hike odds, and gold lost about $170 in two sessions to near $4,150. In lot terms, a standard gold contract is 100 ounces, so those three episodes were worth $5,500, $7,500 of follow-through drift, and $17,000 per lot. The hot CPI was worth $200 per lot in the first minute and $7,500 per lot over three days.

Equities read the same September 11 print the opposite way. The Dow rose about 1% that day to 52,573, because wages were still cooling and the market chose the growth half of the report. It then fell on the hike itself, closing September 16 at 51,462. CPI-day index direction depends on which half wins the first argument, prices or paychecks. Going into Wednesday the roles have swapped: wages are already the cool part at 3.0%, so a hot print argues rates go higher, and equities have less room to call it growth. The Nasdaq historically runs about 1.5 times the Dow's move on these days, in either direction.

Three Scenarios for Wednesday

Probabilities below are my judgment layered on the forecast cluster and market pricing, not consensus numbers. The direction maps matter more than the exact percentages.

The hot case gets 30% for a specific reason. Natixis argues August was the outlier, and the forecast cluster agrees, sitting at 0.16-0.21 core against August's 0.29. But consensus already expects the annual headline at 3.6-3.7%, which would be the hottest annual reading of this hiking cycle, and the energy inputs that drove August (gasoline +3.9%, energy +16.3% year over year) kept rising through September. The consensus path is a hawkish-looking print. The genuine surprise runs through core: a repeat of 0.3% against a 0.2% consensus would reprice December in minutes, and the bank forecasts show almost nobody positioned for it.

The cool case is the violent one because it contradicts 70% December pricing. If core lands at 0.2% or lower and shelter cooperates, the "one more increase" majority loses its evidence two days after the minutes made the case. December odds would fall 20 points fast, the dollar would give back a chunk of its month, and gold would bounce off a tape that just spent a week being sold. First moves overshoot, so into a cool print you take partials on the first vertical candle and trail the rest.

How to Trade It

Before 8:30 AM ET:

• Halve your normal size. Gold spreads widen from cents to over a dollar at the print, so a market order can fill $1-2 worse than the last screen price, which is $100-200 of instant slippage per lot.
• Keep stops outside the event range. The last three repricing episodes traveled $55 to $170 an ounce. A $15 stop during a CPI print is loose change handed to whoever is on the other side.
• Write the decision tree before the print: which core number makes you long, which makes you short, which makes you flat. Add a branch for shelter, since it is the component the doves need. At 8:30:01 you will not have time to think, only to execute decisions you already made.

After the print:

• Read in order: core month over month first, shelter second, the annual headline last. The first two are tradeable in real time. The third is for the evening recaps, not the tape.
• Wait 5-10 minutes for the second move. September 11 is the case study: the spike was worth $200 per lot and the three-day follow-through $7,500. Algorithms trade the headline in milliseconds; the durable move forms once the desks have worked through the details.
• Take partials mechanically at the first target, move the stop to break-even on the remainder, and let the 2-year yield tell you whether the repricing has legs. If the 2-year moves less than about 5 basis points on the print, the gold move probably fades into the close. The Jackson Hole speech moved it 11.8, which is what a real repricing looks like.
• Prop traders: one print points every account the same direction at the same second, and firms that count floating losses breach accounts on the spike, not the close. Wednesday is a correlated-risk day across the whole stack. Size for the worst account, not the best one.
• If you trade IST hours like we do: the release lands at 6:00 PM and the New York open at 7:30 PM IST gives the cleaner, post-algorithm entry. If you cannot watch it live, flat is a position. And remember Thursday 6:00 PM IST brings September PPI, so Wednesday night is not the end of the week's risk.

What the Banks Are Saying

Natixis expects 0.19% core and 0.6% headline, calling August's beat a rise in components that should not repeat, and argues the September hike may turn out to be the only one of this cycle. Deutsche Bank has the street-low core at 0.16%. Morgan Stanley has the street-high headline at 0.62%. The Cleveland Fed's nowcast, updated October 7, puts the annual rates at 3.60% headline and 2.39% core, essentially the consensus.

On the policy side, Goldman Sachs and Barclays both spell out the same path: hold on October 28, hike on December 9. Franklin Templeton's Schulze told Morningstar the Fed stays on hold barring an upside surprise meaningfully above consensus, which is another way of saying the burden of proof sits on inflation to re-accelerate. Notice what that implies for the tape: the consensus print IS the cooling print this month. A 0.2% core against 0.3% last month is a disinflation story that still leaves December alive, and markets will have to decide which half of that sentence to trade.

Frequently Asked Questions

What time is the September CPI release on October 14?

The Bureau of Labor Statistics publishes the September Consumer Price Index on Wednesday, October 14, 2026 at 8:30 AM Eastern, which is 5:30 PM in London and 6:00 PM Indian Standard Time. A few aggregator calendars list October 15; the BLS release schedule itself says October 14. September PPI follows one day later, Thursday October 15, at the same hour.

What is the forecast for September CPI?

FactSet's consensus has headline CPI at 0.5% month over month and 3.6% year over year, with other medians running to 0.6% and 3.7%. Core CPI is forecast at 0.2% month over month, holding at 2.4% year over year. The extremes: Deutsche Bank has the lowest core forecast at 0.16% and Morgan Stanley the highest headline at 0.62%. The Cleveland Fed nowcast, updated October 7, sits at 3.60% headline and 2.39% core for the year. August printed 0.4% headline and 0.3% core.

Why does this CPI matter more than usual?

It is the last inflation reading before the Fed's final rate decision of 2026. The November CPI publishes December 10, one day after the December 9 decision, so the committee votes without it. The September minutes say most officials want one more hike by year-end, and CME FedWatch pricing puts December near 70%. September's weak jobs report (+29,000 against forecasts near 90,000) already removed most of the case for an October move, which funnels the entire year-end debate into this print and Thursday's PPI.

How much does gold move on CPI day?

Recent tape says $55 to $170 per ounce is on the table, which is $5,500 to $17,000 per standard lot. But the September 11 hot print is the cautionary detail: gold closed within $2 of the prior day, then fell about $75 over the following two sessions as hike odds jumped from 56% to 84%. The day-one spike understated the repricing, and the follow-through carried the money. Wait for the second move.

The Short Version

• September CPI lands Wednesday 8:30 AM ET / 6:00 PM IST: consensus 0.5% headline m/m, 3.6-3.7% y/y, core 0.2% m/m holding 2.4% y/y
• It is the last clean inflation read before the December 9 decision, since November CPI publishes a day too late; the minutes want one more hike, December odds run near 70%
• The core forecast cluster (0.16-0.21) sits below August's 0.3, so consensus IS the cooling story, and a repeat of August is the out-of-consensus shock
• Gold spot ~$4,165 after a $50 drift and a 6% September; cool print targets $4,220-4,260, hot extension $4,050-4,090, in line chops $4,130-4,190
• Half size, stops outside the $55-170 event range, second move over first, and PPI follows Thursday 6 PM IST