Core PCE September 30: The Inflation Print That Rewrites Its Own History

September 29, 2026 · 8 min read · ChartWise Team

The Fed's preferred inflation gauge lands Wednesday at 8:30 AM ET (6:00 PM IST), and this month the release comes with a catch. On the same morning, the agency that produces the data rewrites five years of history. August core PCE is forecast at 0.3% month over month after 0.2% in July, which is already enough to move October rate odds. And the number you remember from July, 3.3% year over year, will have been replaced by 8:31 AM.

Gold's Bad Monday Sets the Stage

Before you even look at the forecast, look at what the market did Monday. Gold futures closed Friday near $4,321 and fell more or less in a straight line through Monday's session: $4,230 by 3:00 AM ET, $4,177 by 8:30 AM, a weak bounce to $4,190 around midday, then a fresh low of $4,150 into the New York afternoon. Spot closed around $4,122, down 3.8% on the day per the USA Today price tracker, and futures sat near $4,150 by the Monday evening reopen.

Reuters and Mettis Global attributed the drop to surging Treasury yields and rising odds of another Fed hike, with oil staying elevated because the US-Iran negotiations over the Strait of Hormuz keep going nowhere. The specifics: the 10-year yield reached 5.24%, the highest since 2007 and up 48 basis points in four weeks. The 2-year, the most rate-sensitive instrument there is, touched 4.94%. Traders now price roughly a 65% chance of a 25bp hike on October 28 and about 53% for December. The dollar index held near 101.25, EUR/USD sat around 1.1370, and the Dow slipped 1.1% to 51,482.

Which means the market walked into this PCE print already paying for a hawkish outcome. That changes the surprise math in both directions. A hot print confirms what the tape already says, so part of the move is pre-paid and the continuation may be quieter than the headlines suggest. A soft print contradicts a market that just spent two days pricing the opposite, and contradicted markets move violently. Monday's $170 flush also thinned the bullish positioning that cushioned gold all summer, so there is less baggage to unwind on a bounce.

One more piece of arithmetic worth holding onto. After the September 16 hike to 3.75%-4.00%, the policy midpoint of 3.875% sits about 0.48 points above the 3.4% core PCE rate. That is the first genuinely positive real short rate of this cycle, and it is the Fed's whole argument for restraint. Every 0.1 the core rate climbs eats roughly a fifth of that cushion. A 0.3% print Wednesday probably pushes core year-over-year toward 3.5%, which would shrink the cushion to about 0.38 and make the December hike close to a coin flip that resolves yes.

The August Core PCE Forecast

PCE is the Fed's target measure, and core PCE strips out food and energy the same way core CPI does. It differs from CPI in coverage and weighting, and it lands about ten days after CPI each month, which is why this week feels like a rerun. The inputs that feed it are already public, which is where the forecasts below come from. A nowcast, if you have not run into the term, is a same-week estimate built from those already-published inputs.

Two forecasts bracket the whole trade. TD Securities (Eli Nir and Oscar Munoz) look for 0.24% month over month, 3.3% year over year, after stronger-than-expected PCE-relevant components in the August PPI. Bank of America raised its August core PCE tracking from 0.26% to 0.30% once the hot August CPI landed, and its rates desk had already called that CPI report "all clear for a hike." The Cleveland Fed's nowcast, updated September 25, splits the difference at 0.27% and 3.40% year over year, with headline at 0.34% and 3.78%.

Put the annualization math on the forecast spread, because it is the whole story. July's unrounded 0.246% annualizes to 2.99%. A 0.24% print annualizes to 2.9%. A 0.27% print annualizes to 3.3%. A 0.30% print annualizes to 3.7%. In other words, the gap between the two bank forecasts, six hundredths of a percent, is the gap between "inflation is drifting back toward target" and "December is live." When the smart money's own estimates sit on both sides of the decision line, small prints move markets disproportionately.

Also on Wednesday morning, 75 minutes before PCE: ADP employment at 7:15 AM ET, forecast +73K after +38K. Both feed the same October debate, and a mismatched pair (strong ADP, cool PCE) is a classic chop generator until the second reading settles the argument.

The Revision Trap: 3.3% Stops Existing at 8:31 AM

Here is what makes this release different from every PCE day this year. Per the BEA's own scheduling, the 2026 annual update of the national accounts lands the same morning as the August data, and it rewrites the personal income and outlays series from Q1 2021 through Q1 2026. Three deflators get reworked: portfolio management (fund management fees), legal services, and software. This is not a footnote. The portfolio management line alone supplied roughly half of July's core increase, and its input price (the PPI for portfolio management) jumped 5.39% in July, then fell 1.43% in August on the method now being retired.

The trap is that the comparison everyone will make at first glance is the wrong one. July core PCE was published at 3.3% year over year (3.344% unrounded) on the old method. August's year-over-year rate will be computed entirely on the new one. So:

• If the annual revision carries revised July down to 3.1%, then August at 3.3% is an acceleration, and the "inflation steady" headline is wrong.
• If revised July comes back at 3.4%, then August at 3.3% is a genuine decline dressed up as "unchanged."
• If the print itself lands at 0.2% but the revised history runs hot, the tape will probably drop on the headline and recover as the desks work through the details. That recovery is the trade.

Desks learned this lesson the hard way in past annual updates: the first twenty minutes of a revision-morning tape are algorithmic, and the durable move forms once the re-weighted numbers are understood, usually 15 to 45 minutes later. Plan your decision tree around the month-over-month print and the revised-July headline, and treat the year-over-year number as the last thing worth trusting, not the first.

Why This Print Moves October

Two weeks ago the Fed did something it had not done since 2023: it hiked. The September 16 vote was 12-0, taking the funds rate to 3.75%-4.00%, and the accompanying dot plot put the end-2026 median at 4.1%, which is the Fed's own way of saying one more 25bp move is in the base plan. Warsh's presser language was blunt: inflation has been "too high for too long," and the hike was about removing "a dose of accommodation." His stated standard is confidence "that underlying inflation is moving to our objective clearly and at sufficient speed." Wednesday's print is the first direct test of that sentence.

The market has already voted. About 65% odds on an October hike and 53% on December, per CME FedWatch pricing reported by Trading Economics, with Polymarket's contract around 66%. The 10-year at 5.24% and the 30-year mortgage back above 7% say the bond market is not waiting for the Fed to catch up. Against that backdrop, this is the last inflation reading before the October 27-28 meeting. ISM manufacturing lands Thursday (prior 54.8) and NFP Friday (+90K forecast after +162K), but PCE is the gauge the Fed itself targets, which means it carries more weight per decimal point than either.

The three live paths into the October meeting:

• Core at or below 0.2% with benign revisions: October odds retreat toward 45-50%, "peak rates" is back in the vocabulary, and the December question reopens.
• Core at 0.24-0.28%, revisions neutral: nothing is decided. October stays priced around 60-65% and the meeting becomes a live decision rather than a foregone conclusion.
• Core at 0.3%+ or hot revised history: December locks in above 70%, the dots start looking conservative, and the 10-year tests 5.4%.

The Blueprint: Gold's Last Three Repricing Days

The honest way to size Wednesday's risk is to look at what gold actually did on its last three event days, because the range is wider than most traders assume.

August 12, in-line CPI: gold ran from $4,381 to a $4,438 high, about $55 an ounce (1.2-1.5%), on a print that surprised nobody. September hike odds fell from 46% to 40% and the 2-year dropped 4bp. August 28, Jackson Hole: Warsh called inflation "too high for too long" and gold lost $70 in roughly ten minutes ($4,626 to $4,554), extending to a 3.14% loss by the close near $4,456. The 2-year rose 11.8bp to 4.348%, the dollar index gained 0.5%, and September hike odds went from 35.9% to 57.5% in a session. September 28-29, no data print at all: yields and hike odds alone took gold down about $170 to near $4,150.

In lot terms, a standard gold contract is 100 ounces, so those three days were worth $5,500, $7,000, and $17,000 per lot. Even a 0.10 lot position, the "small" size many traders start funded accounts with, swung $550 to $1,700. For comparison, a 100-pip EUR/USD day, a large one, is worth $1,000 per standard lot. Gold on an event day moves like three currency pairs at once, which is why we keep saying CPI and PCE days are gold stories first and forex stories second.

The lesson from the three days is that gold trades the change in the expected rate path, not the calendar. An in-line print moved $55 because positioning was nervous going in. A speech moved $170 because it changed October. PCE can do either, and this month the revision wildcard means even an in-line print can carry an agenda.

Three Scenarios for Wednesday

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

Scenario 3 gets 40% for a reason. Normally you would cap the hot case near 25%, since the consensus sits at 0.3% anyway and BofA's tracking is already there. But the market front-ran the outcome on Monday: a hot print now mostly confirms pricing, and confirmation days still trend, they just extend rather than gap. The nastier version is the quiet one: a 0.25% headline with revised history showing acceleration. That combination reads "in-line" in the first minute and "hawkish" by 9:00, and the second move is the one that pays.

Scenario 1 is the violent one precisely because it is out of consensus. TD at 0.24% is the loneliest forecast on the street right now, and if it wins, October odds drop 15-20 points in an hour, the dollar gives back a week of gains, and gold bounces off the most oversold tape since the January nomination shock. First moves overshoot, so into a soft 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. Revision-morning tape can whipsaw twice before it commits, and both whipsaws cost real money. 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 recent event-day range. The last three event days traveled $55 to $170 an ounce. A $15 stop during PCE is a donation to the liquidity providers.
• Write the decision tree down before the print: which number makes you long, which makes you short, which makes you flat. Include the revised-history headline as a separate branch. At 8:30:01 you will not have time to think, only to execute decisions you already made.

After the print:

• Read in order: month-over-month core first, the revised July second, the year-over-year last. The first two are tradeable in real time. The third is a headline for people who read news after lunch.
• Wait 5-10 minutes for the second move. Algorithms trade the headline in milliseconds; humans trade the revisions over the next half hour. The classic pattern is spike, reverse, trend, and the trend is the part you can actually join.
• Take partials mechanically at the first target (pre-release high or low), move the stop to break-even on the remainder, and let the 2-year yield tell you whether the rate repricing has legs. The 2-year moved 11.8bp on Jackson Hole day; if it moves less than 5bp Wednesday, the gold move probably fades into the close.
• Prop traders: the print hits at 8:30 AM ET, and firms that count floating losses against a daily limit breach accounts on the spike, not the close. If you run multiple funded accounts, Wednesday is a correlated-risk day, since one print points every account the same direction at the same second. Know each firm's reset clock before you size up anywhere.
• 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 the print live, flat is a position, and Friday's NFP gives the week a second chance to pay you.

What the Banks Are Saying

TD Securities: "We now look for 0.24% m/m for August core PCE (3.3% y/y) with a stronger 0.33% for the headline (3.8% y/y)," noting that PCE-relevant components of August PPI came in slightly stronger than anticipated and flagging the BEA methodology change, with an expectation of downward revisions to year-over-year inflation but a firmer near-term monthly profile.

Bank of America Global Research raised its August core PCE tracking from 0.26% to 0.30% after August CPI, the same report their rates desk called "all clear for a hike" two weeks before the Fed delivered one. The Cleveland Fed's nowcast model, updated September 25, puts core at +0.27% m/m and 3.40% y/y. Forex Factory's consensus line reads 0.3%.

Notice where the bank cluster sits: 0.24 to 0.30, with the decision line at roughly 0.25. When the forecasts hug the rounding boundary this tightly, a four-hundredth miss in either direction reprices October, and the desks know it. Thin consensus is exactly how a "boring" 0.27% print still moves gold $40.

Frequently Asked Questions

What time is the core PCE release on September 30?

The Bureau of Economic Analysis publishes the August Personal Income and Outlays report on Wednesday, September 30, 2026 at 8:30 AM Eastern, which is 5:30 PM in London and 6:00 PM Indian Standard Time. Forex Factory carries a red (high-impact) flag on the core reading, and the same release includes personal income, personal spending, and the final second-quarter GDP estimate.

What is the forecast for August core PCE?

Forex Factory shows 0.3% month over month against 0.2% in July. The Cleveland Fed nowcast, updated September 25, puts August core PCE at +0.27% and 3.40% year over year. TD Securities forecasts 0.24% and 3.3%, and Bank of America raised its tracking estimate to 0.30% after a hot August CPI. The rounding line sits near 0.25%: prints at or below it annualize under 3.3%, prints above it run hotter.

Why does the PCE revision matter for traders?

The BEA 2026 annual update rewrites the index from Q1 2021 through Q1 2026 and lands the same morning as the August data. July core PCE was published at 3.3% on the old method, and that number stops existing at 8:31 AM. The only valid comparison is August against the revised July inside the same release. A drop in the annual rate could be a method change rather than cooling, and a flat print could hide a real acceleration if the history was revised down.

How much does gold move on PCE day?

Recent tape says $55 to $170 per ounce is on the table. Gold ran about $55 (1.2-1.5%) on an in-line CPI print on August 12, and it fell roughly $170 (about 3.9%) on Monday, September 29 after yields surged and October hike odds hit about 65%. A standard gold lot is 100 ounces, so those moves are $5,500 and $17,000 per lot.

The Short Version

• August core PCE lands Wednesday 8:30 AM ET / 6:00 PM IST, forecast 0.3% m/m after 0.2%, with the Cleveland Fed at 0.27% and TD at 0.24%
• The BEA rewrites the PCE index back to 2021 the same morning, so compare August only to the revised July in the same release
• The annualization line: 0.24% is 2.9% a year, 0.30% is 3.7% a year, and that spread is the gap between a pause debate and a locked December hike
• Gold already fell $170 into the print; the soft scenario is now the surprise, so it is the violent one. Hot extension targets $4,050-4,080, soft bounce $4,220-4,250
• Half size, stops outside the $55-170 event range, second move over first, and remember NFP (+90K forecast) is Friday