The New York Fed Calls It a Four-Year High. In the Same Table, the Orders and Shipments Indexes Both Fell.
The Federal Reserve Bank of New York published its August Empire State Manufacturing Survey at 8:30 a.m. Eastern on Monday, and the headline did what headlines are supposed to do: it moved. The general business conditions index rose to 20.6 from 15.6 in July, a five-point gain that the New York Fed described in its own summary as "its highest reading in more than four years." The wire services had it inside a minute. InvestingLive, which timestamped its post at 12:30:31 UTC — 8:30 a.m. in New York — measured the print against an estimate of 11.00. Investing.com, covering the same release, put the forecast at 10.60. Those two providers do not agree, so this article adopts neither as the consensus; on either number the print landed roughly ten points above what that provider had published.
That is the story most readers will get. It is accurate, and it is incomplete. The August report is a two-page document with tables underneath it, and the tables say something more complicated than the headline does. The indexes for two of the three indicators a reader would most likely assume sit behind a business-conditions number — new orders and shipments — fell in August. The third, the number of employees index, fell too. All three stayed above zero, and the New York Fed's own write-up says orders and shipments "posted solid gains" and that "employment levels and the average workweek rose modestly." What moved down was the index, which measures how widely the growth was reported, not the direction of activity.
Start with how the survey is constructed, because the arithmetic matters. Empire State is a diffusion survey. Each month the New York Fed sends the questionnaire to roughly 200 manufacturing executives across New York State and typically gets about 100 back, according to the bank's methodology page. Respondents are asked whether each indicator rose, fell, or held steady versus the prior month. The published table for current indicators carries exactly three columns of figures, under a "Percent Reporting" spanner, labeled "Higher," "Lower" and "Index." There is no "no change" column, and the release carries a footnote reading "Data are seasonally adjusted." The index is the first column minus the second: for shipments in August, 37.4 percent higher minus 25.7 percent lower gives 11.7 exactly. The identity reproduces most rows on the nose and the rest within a tenth — general business conditions prints 43.8 percent higher and 23.1 percent lower against a published index of 20.6, where the rounded shares would give 20.7 — because the index is computed from unrounded shares. Each indicator's table shows two months, July and August, plus a change row. Every current-conditions figure below comes from that table; the six-months-ahead figures, flagged as such, come from the separate forward-looking table. Nowhere in this two-page release does the New York Fed print an unadjusted counterpart to either.
Now the components. The new orders index printed at 17.3, down from 22.2 in July. The shipments index printed at 11.7, down from 24.4. That 12.7-point drop in shipments is the largest single move anywhere in the current-conditions table this month, larger than the headline's own five-point gain and larger than any other component's swing. The number of employees index eased to 9.3 from 11.4. None of those three went negative — a positive diffusion index still means more firms reported growth than contraction, so orders, shipments and headcount were all still expanding in August. What narrowed was the breadth. Fewer firms reported higher shipments in August (37.4 percent) than in July (45.3 percent), and more reported lower ones (25.7 percent versus 21.0 percent).
So how does the headline rise while its apparent inputs fall? Because they are not its inputs. The New York Fed states this plainly on the survey's overview page, under the heading "About Survey": "The survey's main index, general business conditions, is not a weighted average of other indicators—it is a distinct question posed on the survey." It is its own line on the questionnaire. An executive can report fewer shipments than last month and still answer that general conditions at the firm improved, and in August a meaningful number of them did exactly that. Responses were collected between August 3 and August 10, per the report.
The indicators that surged are clustered on the supply and cost side of the questionnaire rather than the demand side. The unfilled orders index rose to 15.5 from 5.0, a 10.5-point gain that is the largest increase in the current table. The delivery times index rose to 20.6 from 13.0, meaning substantially longer lead times. The supply availability index fell further into negative territory, to -13.4 from -10.0. Inventories flipped from +4.0 to -5.2, the only negative inventories reading in the two months this release prints.
Read together, that is a specific and recognizable pattern. Order backlogs building, deliveries slowing, inventories drawing down and supply getting harder to source is what a constrained factory sector looks like. It is not the same thing as a demand acceleration, and it can coexist comfortably with the shipments index falling by nearly thirteen points — firms that cannot get inputs ship less even when their order books are fine.
The price columns point the same direction. The prices paid index rose to 58.6 from 52.3. The composition is the striking part: 59.4 percent of respondents reported paying higher prices in August and 0.8 percent reported paying lower ones. Fewer than one respondent in a hundred saw input costs fall. Prices received went the other way, easing to 22.7 from 27.6, which the New York Fed characterized as selling price increases that "remained elevated but eased for a second consecutive month." Input costs accelerating while output prices decelerate is a margin squeeze in the current month, whatever it eventually means for consumer inflation.
A year-ago comparison sharpens it. In the New York Fed's September 2025 report, which shows August and September 2025 side by side, the August 2025 general business conditions index was 11.9, prices paid was 54.1, number of employees was 4.4, delivery times was 17.4 and supply availability was -5.5. Against that base, this August's headline is roughly nine points better, input-price pressure is 4.5 points higher, and supply availability is nearly eight points worse. The cost side and the supply side have both deteriorated over twelve months even as the sentiment question improved.
The forward-looking table — the same three columns, six months ahead — is where the survey's optimism actually lives. Future general business conditions rose to 32.1 from 27.9. Future new orders reached 37.1 and future shipments 33.7. The two biggest moves in the entire forward table are future unfilled orders, which swung to 19.6 from -3.0, a 22.6-point jump, and future employment, which rose to 28.2 from 14.4. Capital expenditure plans barely moved, to 16.5 from 15.0; the report's own summary says "Capital spending plans remained modest." Notably, future prices received climbed to 48.7 from 41.9 while future prices paid rose to 57.7 from 53.0: firms expect to be paying more six months out, and they expect to pass more of it along than they are managing to now.
The limits are worth stating. This is one state, roughly a hundred responses, and a diffusion measure that counts firms rather than dollars — a large manufacturer and a small one each cast one vote, and the index says nothing about the magnitude of any firm's change. Regional Fed surveys are noisy month to month, which is precisely why a five-point headline move deserves the components rather than a standalone reading. Empire State is also the first of the August regional manufacturing prints, and it has diverged from the others before.
It arrives in a specific policy window. The Federal Open Market Committee's next scheduled meeting is September 15-16, one of the four 2026 meetings the Fed's calendar marks with an asterisk for a Summary of Economic Projections. Nothing in a single regional survey settles that meeting. But the August Empire State table hands both sides of the internal argument something: a headline the New York Fed itself calls "a more than four year high" for anyone arguing activity is holding up, and a prices paid index at 58.6 with almost no firms reporting cost relief for anyone arguing the inflation impulse has not finished working through the goods pipeline.
Monday's trading session was still underway and unfinished as this was written, and the survey is the sort of second-tier release that rarely reprices a whole tape on its own. The document itself, though, is fully public and takes about ten minutes to read column by column. Anyone repeating the four-year-high line should be able to say what happened to shipments in the same table.
Sources & further reading
- Federal Reserve Bank of New York, "Empire State Manufacturing Survey — August 2026", released August 17, 2026, accessed August 17, 2026
- Federal Reserve Bank of New York, "Empire State Manufacturing Survey (overview)", "About Survey" section, accessed August 17, 2026
- Federal Reserve Bank of New York, "Economic Indicators Calendar" (Empire State Manufacturing Survey, Monday, August 17, 2026, 08:30), accessed August 17, 2026
- Federal Reserve Bank of New York, "Empire State Manufacturing Survey — September 2025", September 2025, accessed August 17, 2026
- InvestingLive, "US Empire Manufacturing index 20.60 vs 11.00 estimate", August 17, 2026, accessed August 17, 2026
- Board of Governors of the Federal Reserve System, "FOMC Meeting calendars and information", accessed August 17, 2026
- Investing.com, "Empire State Manufacturing Index Surges, Surpassing Expectations", August 17, 2026, accessed August 17, 2026