Philly Fed 37.8 vs 31.3 Est — Firms' Own Price Expectations Highest Since 1981
Fundamentals · 2026-09-17
The Philadelphia Fed's general activity index fell 9.6 points to 37.8 in September, against a 31.3 consensus — a 6.5-point beat, and still the 29th highest reading in 701 months since 1968. Demand barely moved: new orders 29.2 from 30.1, shipments unchanged at 27.7. The prices went the other way — prices paid 48.6 (+7.7) and prices received 31.3 (+13.6), the latter the highest since April. Forward, the number that matters: future prices received rose 12.5 points to 72.3, its highest reading since September 1981. Future prices paid 71.3. Delivery times jumped 22.8 points to 26.5, the second-largest monthly move in the survey's history. Employment fell 16 points to 11.8, almost entirely firms moving to "no change". 72% now call labor supply a constraint, up from 50% in June. Inventories -12.5. Survey window 7-15 September.
What It Changes
- Firms expect to raise their own prices more aggressively than at any time since 1981. The future prices received index at 72.3 ranks 33rd of 701 months, and the last month above it was September 1981 — forty-five years ago, mid-Volcker. That is not a cost complaint; it is a statement about pricing power.
- It answers the question this desk posed on Monday, and the answer is split. After Empire, we asked whether the New York cost spike was regional or general. Costs are general — both districts show prices paid and forward prices paid rising. The demand stall is not. New York's new orders fell to 2.0 and shipments went negative; Philadelphia's new orders held at 29.2 and shipments did not move at all.
- Supply chains are seizing again. Delivery times jumped 22.8 points, the second-largest monthly increase in 58 years, behind only March of this year. Two of the largest lengthening episodes on record have happened six months apart, in 2026.
- The headline fell and beat at the same time, which is the correct read. Down 9.6 points from a five-year high, still 6.5 above consensus, and still in the 96th percentile of the survey's history. Philadelphia's 2026 average is 23.4 against 4.6 for 2025.
- It gives the Fed exactly what it said it was worried about, one day late. Warsh justified Wednesday's hike on "too many categories posting increases above 3%" and a commitment to stop price increases broadening out. This is a second district reporting accelerating pass-through.
Impact
- USD — Slight bullish — a survey beat with forward price expectations at a 45-year high supports the path the Fed published on Wednesday, though Philadelphia rarely moves the dollar alone.
- DXY comes in having tagged 100.00 during the decision and closed up 0.35% at 99.97, with the 10-year back above 5% at 5.009%. Nothing here argues against those levels.
- US Indices (ES / NQ / YM / RTY) — Mixed, lean bearish — activity beating is good for earnings, but a 45-year high in expected selling prices is the input that keeps the October meeting live.
- The margin read is more useful than the index read. Prices received rose 13.6 points against prices paid up 7.7, so Philadelphia firms are passing costs on faster than they are absorbing them. That is better for manufacturer margins than the Empire picture, where the paid-minus-received gap sits at 35 points.
- Futures came in bid overnight — ES +0.85%, NQ +1.10%, YM +0.72%, RTY +0.79% — but on 13% to 18% of average volume, ahead of Friday's triple witching.
- Gold (GC) — Slight bearish — gold is trading as a rates instrument, and forward price expectations at 1981 levels push real-rate expectations the wrong way for it.
Inside The Number
The diffusion index for current general activity fell 9.6 points to 37.8 in September from August's 47.4, against a consensus of 31.3. So the index declined sharply and beat expectations by 6.5 points, and both statements are true and worth holding together.
Start with where 37.8 actually sits, because the decline is misleading. The survey has run monthly since May 1968 — 701 observations. A reading of 37.8 ranks 29th, meaning 96% of months in the survey's history have been weaker. The full-history average is 8.9. Philadelphia's 2026 average is 23.4, against 4.6 for 2025. This district is having an exceptional year and September did not change that.
August's 47.4 was, as billed, a five-year high: the last month above it was April 2021, at 48.8. Falling 9.6 points from a five-year high still leaves you in the top 5% of readings ever recorded.
Demand did essentially nothing, which is the most important thing in the release. New orders edged down 1 point to 29.2. Shipments were unchanged at 27.7. Unfilled orders rose 6 points to 20.4. Over 44% of firms reported higher new orders against 15% reporting decreases. There is no demand deterioration in this survey at all.
Now the prices, which is where the release earns its place. The prices paid index rose 7.7 points to 48.6, recovering most of last month's 13-point fall. Nearly 51% of firms reported higher input costs, up from 41%, and only 2% reported decreases. The prices received index rose 13.6 points to 31.3, its highest since April — 33% of firms raised their own prices, up from 21%.
Note the ratio. Prices received rose almost twice as fast as prices paid. In Philadelphia, pass-through is not just happening, it is accelerating faster than the cost pressure behind it.
And the forward price expectations are the finding. The future prices received index rose 12.5 points to 72.3. Against 701 months of history that ranks 33rd, and the last month with a higher reading was September 1981 — when the funds rate was in double digits and Volcker was two years into breaking the great inflation. Forty-five years. The future prices paid index rose 8.4 points to 71.3, itself in the 80th percentile.
Read those two together and the message is unambiguous: Philadelphia manufacturers expect input costs to keep rising *and* expect to be able to push them through to customers. That combination is the definition of an environment where inflation broadens, and it arrived the day after a Fed Chair said "we will ensure any changes in prices won't broaden out."
Delivery times produced the largest single move in the report and almost nobody will mention it. The index jumped from 3.7 to 26.5, a rise of 22.8 points — the second-largest monthly increase in 700 months of the survey, behind only March 2026's +28.8. The level itself ranks 9th of 701, a 99th-percentile reading, and the last month higher was March 2022, at the peak of the post-pandemic supply-chain crisis. The future delivery times index went from 0.0 to 18.6, so firms expect the lengthening to continue.
Two of the largest supply-chain lengthening episodes in the survey's entire history have now happened within six months of each other, both in 2026. Set that beside the Empire survey's delivery times at 18.8 and supply availability at -11.9, and the import price data showing nonfuel industrial supplies up 12.6% on the year, and this is not a Philadelphia story.
Employment fell hard, and the composition is the usual reassurance. The index dropped 16.1 points to 11.8, mostly undoing August's increase. But the firms reporting increases fell from 33% to 17.3% while those reporting decreases rose only from 5% to 5.5%. The residual went to "no change," which climbed from 62% to 77.1%. That is hiring pausing, not firing starting — the same pattern Empire showed on Monday, the same pattern NAHB showed yesterday, and entirely consistent with this morning's claims print at 196,000.
The average workweek index fell 8.5 points to 18.0, which the release calls a third consecutive month of elevated readings. It still ranks 34th of 701.
Inventories fell to -12.5 from -3.7, the lowest since January 2024, so Philadelphia firms are drawing down stock. That is the opposite of Empire, where inventories swung positive to 8.9, and the opposite of the national picture where wholesale inventories are up 5.7% on the year.
Forward expectations softened but from an extraordinary level. Future general activity fell 20.7 points to 52.9, the largest forward decline in the release — yet 52.9 still ranks 160th of 701 and sits well above the survey's history. Future employment actually *rose* 15.2 points to 50.6 and the future workweek rose 10.4 to 31.8. Only future capital expenditures went meaningfully the wrong way, down 11.1 points to 37.1.
Against Empire, the comparison is the most useful thing in this brief. Both surveys cover September; Empire collected 2-10 September, Philadelphia 7-15 September, so Philadelphia's window includes Friday's CPI and the Saudi supply escalation that Empire's did not.
The Internals
Current indexes, September against August:
Index · September · August · Change · Read
General activity · 37.8 · 47.4 · -9.6 · Fell hard, beat consensus by 6.5, 96th percentile
New orders · 29.2 · 30.1 · -0.9 · Demand essentially unchanged
Shipments · 27.7 · 27.7 · 0.0 · Literally unchanged
Unfilled orders · 20.4 · 14.4 · +6.0 · Backlogs building
Delivery times · 26.5 · 3.7 · +22.8 · Second-largest monthly jump in 58 years
Inventories · -12.5 · -3.7 · -8.8 · Lowest since January 2024
Prices paid · 48.6 · 40.9 · +7.7 · Recovering last month's fall
Prices received · 31.3 · 17.7 · +13.6 · Highest since April; rose twice as fast as paid
Employment · 11.8 · 27.9 · -16.1 · Hiring paused, not reversed
Average workweek · 18.0 · 26.5 · -8.5 · Third elevated month
Expectations six months ahead:
Index · September · August · Change · Read
Future general activity · 52.9 · 73.6 · -20.7 · Largest forward decline, still 77th percentile
Future new orders · 62.3 · 66.0 · -3.7 · Demand expectations intact
Future shipments · 61.1 · 63.5 · -2.4 · Intact
Future unfilled orders · 16.8 · 8.5 · +8.3 · More backlog expected
Future delivery times · 18.6 · 0.0 · +18.6 · Lengthening expected to continue
Future inventories · 11.6 · 17.4 · -5.8 · Less restocking planned
Future prices paid · 71.3 · 62.9 · +8.4 · 80th percentile
Future prices received · 72.3 · 59.8 · +12.5 · Highest since September 1981
Future employment · 50.6 · 35.4 · +15.2 · The largest forward increase
Future workweek · 31.8 · 21.4 · +10.4 · More hours expected
Future capital expenditures · 37.1 · 48.2 · -11.1 · The one forward index that fell meaningfully
Where this month's standouts sit in 701 months since May 1968:
Reading · Value · Rank · Last month higher or lower
Delivery times · 26.5 · 9th highest, 99th percentile · March 2022, at 37.9
General activity · 37.8 · 29th highest, 96th percentile · August 2026, at 47.4
Future prices received · 72.3 · 33rd highest, 95th percentile · September 1981, at 78.5
Average workweek · 18.0 · 34th highest, 95th percentile · August 2026, at 26.5
Prices received · 31.3 · 90th highest, 87th percentile · April 2026, at 33.5
Future prices paid · 71.3 · 140th highest, 80th percentile · November 2025, at 74.5
Prices paid · 48.6 · 143rd highest, 79th percentile · July 2026, at 53.9
Inventories · -12.5 · 533rd, only 24% of months lower · January 2024, at -16.0
Philadelphia against New York, the same month, two panels:
Index · Empire, New York · Philadelphia · Read
General activity · 7.6 · 37.8 · Two very different districts
New orders · 2.0 · 29.2 · The demand stall is regional
Shipments · -3.2 · 27.7 · New York negative, Philadelphia unchanged
Prices paid · 63.1 · 48.6 · Both elevated, both rising
Prices received · 28.1 · 31.3 · Both rising; Philadelphia faster
Future prices paid · 67.3 · 71.3 · Both near multi-year highs
Future prices received · 42.3 · 72.3 · The sharpest divergence in the pair
Employment · 10.6 · 11.8 · Almost identical
Average workweek · 17.0 · 18.0 · Almost identical
Future general activity · 29.0 · 52.9 · Philadelphia far more optimistic
2026 by month, general activity:
Month · Index · Month · Index
January · 12.6 · June · 10.3
February · 16.3 · July · 41.4
March · 18.1 · August · 47.4
April · 26.7 · September · 37.8
May · -0.4 · 2026 average · 23.4
What The Firms Actually Said
The special questions this month covered production and capacity, and one answer stands out. Asked to compare third-quarter production with the second quarter, 68% of firms reported an increase against 16% reporting a decrease — with 12% reporting a rise of 10% or more. That is a strong quarter by any reading, and it corroborates the activity index rather than the headline decline.
Capacity utilization did not move. The median reported rate was 70-80% both this quarter and a year ago. The distribution shifted slightly upward at the top — 16.7% of firms now report 90-100% utilization against 12.0% a year ago, and 12.5% report 80-90% against 8.0% — but the middle of the distribution is unchanged. Manufacturers are producing more without running materially hotter.
The labor constraint is the number that jumped, and it jumped a long way. Asked what is constraining capacity utilization, 72% of firms named labor supply as at least a slight constraint, up from 50% when the question was last asked in June. The breakdown: 28% not at all, 28% slightly, 32% moderately, 12% significantly. A 22-point increase in a single quarter is a fast deterioration in labor availability, and it sits alongside NAHB's chairman reporting yesterday that "increased immigration enforcement is discouraging legal workers from reporting to job sites" and Empire's "persistent labor shortages."
That constraint also reframes the employment index. Firms reporting no change in headcount rose to 77.1%. If nearly three-quarters of the panel says labor supply is constraining them, a flat employment index may be a supply problem rather than a demand one — firms not hiring because they cannot find people, not because they do not want them.
Other constraints are mild by comparison. Financial capital constrains 84% of firms not at all — notable with the funds rate at 3.75%-4.00% and the 10-year above 5%. Supply chains constrain 40% not at all, though 28% call it moderate. Uncertainty constrains 32% not at all, down markedly.
Looking three months out, energy is the worry. 36% of firms expect the impact of energy markets to worsen, against 4% expecting improvement — the most negative forward answer in the whole battery, and similar to last quarter. Uncertainty is the one that improved: only 20% expect it to worsen, down from 44% last quarter. Labor supply and supply chains are each expected to worsen by 16%.
With crude above $100, diesel at a record and the Saudi East-West pipeline still only half-restored, 36% of firms flagging energy as a deteriorating constraint is the mechanism by which this quarter's supply shock reaches next quarter's prices.
On the distributions behind the headline indexes, the general activity split was unusually balanced: equal shares of firms — 45% each — reported increases and no change, with only 8% reporting decreases. A 37.8 index built on 8% negative responses is a much healthier number than one built on a large minority reporting deterioration.
Against This Morning's Open
- The Open called this the fulcrum of the 08:30 block and was right to — The Open wrote that "the Philadelphia Fed has the most room to move inside it," noting a survey "that has just risen far and fast enough to revert hard in either direction."
- It framed the consensus correctly as a base case rather than a forecast. The Open described the expected 16-point decline to 31.3 as "taken as the base case." The actual decline was 9.6 points to 37.8 — a beat, and the reversion was milder than the street assumed.
- Its prior figures check out. August 47.4, July 41.4, June 10.3 all match the Philadelphia Fed's own history file, and the "five-year high" description of August is correct: the last month above 47.4 was April 2021.
- The one thing the setup missed is where the news actually was. The Open framed this release as an activity story. The activity number beat and mattered less than the price indexes, and the 45-year high in expected selling prices is not something a consensus figure of 31.3 was ever going to capture.
What This Sets Up
- This closes the question Monday's Empire brief left open. That brief asked whether the New York cost spike was regional or general. It is general on costs and regional on demand — Philadelphia confirms rising prices paid, rising prices received and rising forward price expectations, but shows none of New York's order stall.
- Next release — Thursday 15 October, covering October, the first survey window fully after the Fed's hike and the first to capture whether the energy shock firms are flagging actually lands.
- Whether future prices received holds above 70 is the single number to watch. At 72.3 it is at levels not seen since 1981. A second month there would make expected pass-through, rather than input costs, the center of the inflation problem.
- Whether delivery times stay above 25. Two of the largest lengthening episodes in 58 years have hit in 2026. A third consecutive elevated reading would mean the supply-chain normalization of 2023-24 has genuinely reversed.
- Whether the labor constraint keeps climbing. From 50% to 72% in one quarter is fast. The question is next asked in December, but the employment index each month will show whether firms keep moving to "no change."
What Is This?
- What it is: The Manufacturing Business Outlook Survey, run by the Federal Reserve Bank of Philadelphia since May 1968 — the longest-running regional manufacturing survey in the United States, with 701 monthly observations. It polls manufacturers in the Third Federal Reserve District, covering eastern Pennsylvania, southern New Jersey and Delaware, and asks each firm whether a measure has increased, decreased or stayed the same versus the prior month. Each index is the share reporting increases minus the share reporting decreases, seasonally adjusted, so zero is the line between expansion and contraction. The same questions are asked about expectations six months ahead, and each quarter the survey adds special questions on a topic of current interest. This month's covered production growth, capacity utilization and the constraints on it. Responses were collected 7-15 September.
- Why it matters: It is the second regional manufacturing survey of each month, landing two days after Empire, which makes it the first genuine cross-check on whether a New York signal is national. It also has the longest history of any of them, so its readings can be ranked against nearly six decades rather than two. Right now its price indexes carry unusual weight: the Fed has just hiked while explicitly worrying about inflation broadening, and this survey asks firms directly what they expect to charge.
- How to read it: Four things. It is a diffusion index, so the number measures how *widely* something is happening, not by how much — 48.6 on prices paid does not mean costs rose 48.6%. It is one district and roughly 100 firms, so month-to-month swings of ten points or more are routine and a single month means very little. The level matters more than the change: a 9.6-point fall that leaves the index in the 96th percentile of its own history is not a deterioration in any meaningful sense. And the forward indexes are opinions, not orders — future prices received is what firms hope to charge, which is informative about intent and says nothing about whether customers will pay.
Want the full explainer? Economic News Events, Explained breaks down this release and every other one we cover, in plain English.
_For informational purposes only. Not investment advice._