The headline is a transportation story until proven otherwise
On October 27, 2026, the Census Bureau released its advance report on durable goods manufacturers’ shipments, inventories, and orders for September. The headline number was strong: new orders rose $6.3 billion, or 1.8 percent, to $345.4 billion. That is the figure most wire stories led with.
Now read the next line of the same release. Excluding transportation, new orders rose 0.5 percent. Transportation equipment alone accounted for $5.2 billion of the $6.3 billion increase, rising 4.6 percent to $119.8 billion.
That is the whole lesson in two sentences. The aggregate moved because one industry group moved. If you want to know whether business demand broadened in September, the headline does not answer the question. The ex-transportation line does, and it says something much quieter.
This is not a flaw in the release. It is a feature of the thing being measured. Durable goods are expensive, long-lived items, and a single large order for aircraft can register as a multi-billion-dollar monthly change. The Census Bureau publishes the components precisely so readers can decompose the total. The failure mode is editorial: reporting the aggregate and skipping the decomposition.
Decomposing the September 2026 release
The advance report’s Table 1 gives the arithmetic directly. Total durable goods new orders were $345,371 million in September, up from a revised $339,117 million in August. The $6.3 billion gain is the difference.
Within that total:
- Transportation equipment new orders rose $5.2 billion to $119,806 million.
- Excluding transportation, new orders rose 0.5 percent, to $225,565 million.
- Excluding defense, new orders rose 2.2 percent.
Transportation is not a monolith either. Motor vehicles and parts new orders rose just 0.3 percent, to $73,445 million. The action was in aircraft. Nondefense aircraft and parts new orders rose 28.4 percent in September, to $25,337 million, after a flat August. Defense aircraft and parts rose 5.7 percent.
So the chain of custody for the headline runs: total durable goods, then transportation, then aircraft, then nondefense aircraft. Each step narrows the story. By the time you reach nondefense aircraft, you are looking at a series that can swing by double digits in a month on the timing of a few large orders.
Why one industry group can dominate the aggregate
The M3 survey measures value, not counts. A commercial aircraft order is one of the largest single manufactured items in the U.S. economy, and orders are booked when they are placed, not when they are delivered. A handful of orders in a month can move the aircraft series by billions of dollars, and the aircraft series is large enough relative to the total to move the aggregate.
The Census Bureau’s own documentation of aggregate series makes the composition explicit. The nondefense capital goods aggregate (series NDE) is built from a defined list of M3 industry series, including nondefense aircraft manufacturing (36F). The ex-aircraft variant (NXA) is defined as nondefense capital goods less 36F. That is a mechanical subtraction, not a judgment call. The Bureau publishes the ex-aircraft line because it knows the aircraft component is lumpy.
This is the practical meaning of the phrase “core capital goods” as it is used in commentary on this release. It refers to the M3 aggregate for nondefense capital goods excluding aircraft. It is a Census Bureau series, not a Federal Reserve or BEA measure of business investment. The distinction matters: the M3 series measures orders and shipments in current dollars, while the business investment components of GDP are measured differently and published on a different schedule. Treating the core capital goods line as a proxy for GDP investment is a common shortcut, and it is a shortcut, not an identity.
The core capital goods line in September
Nondefense capital goods excluding aircraft new orders rose 0.5 percent in September, to $88,039 million. That is the same 0.5 percent as the ex-transportation line, which is not a coincidence: aircraft is the dominant volatile component of transportation, and removing it removes most of the month-to-month noise.
Compare the three numbers side by side:
- Headline durable goods new orders: +1.8 percent.
- Excluding transportation: +0.5 percent.
- Nondefense capital goods excluding aircraft: +0.5 percent.
The headline and the core line tell different stories about the same month. One says orders jumped. The other says orders grew modestly. Both are true. They answer different questions. The headline answers “how much did total durable goods orders change?” The core line answers something closer to “how much did business equipment demand change, setting aside the lumpiest component?”
For readers trying to gauge the direction of business investment, the second question is usually the more useful one. For readers trying to understand why the headline moved, the first question is the right one, but only if you then decompose it.
What the survey mechanics tell you about precision
The advance report is preliminary, and the M3 survey has design features that limit how precisely any single month’s change should be read.
The survey is not based on a probability sample. The Census Bureau states this directly: because the estimates are not based on a probability sample, sampling error cannot be measured and confidence intervals cannot be computed. The panel comprises approximately 4,700 reporting units representing approximately 3,000 companies. Those companies are weighted to represent the universe, but the month-to-month movement of a non-probability panel is an indication, not a measurement with a known error band.
The figures are adjusted for seasonality but not for inflation. A 1.8 percent increase in nominal orders is not a 1.8 percent increase in real orders. Over short horizons the difference is usually small, but over a year it is not, and the release’s year-to-date comparisons are nominal.
The advance report is also subject to revision. The release itself notes that revised and more detailed estimates, plus nondurable goods data, are published in the full report about a week later. The August figures in the September advance report were themselves revised from the prior month’s publication. This is normal. It is also a reason not to treat the advance headline as a final number.
None of this means the release is unreliable. It means the release is a preliminary signal from a voluntary panel, and the appropriate response to a surprising month is to check the components and wait for the full report before drawing strong conclusions.
The “Boeing effect” framing, handled carefully
Commentary on durable goods often reaches for a single company to explain a swing in the aircraft series. The temptation is understandable: commercial aircraft manufacturing in the United States is concentrated, and a production pause or a large order at one manufacturer can plausibly show up in the nondefense aircraft series.
But the M3 release does not attribute orders to named companies. It publishes industry aggregates. The nondefense aircraft and parts series (NAP) is built from industry categories 36F and 36H, and the release does not tell you which firm booked which order in which month. Any claim that a specific month’s swing was “caused by Boeing” is an inference from outside the release, and it should be labeled as such.
That does not make the inference useless. It makes it a hypothesis to be checked against the data rather than a fact to be asserted. If you want to connect a company event to an M3 month, the defensible procedure is: identify the event and its timing from the company’s own disclosures, check whether the nondefense aircraft series moved in the corresponding month, and state the connection as a possible contributor rather than a demonstrated cause. The release gives you the series; it does not give you the attribution.
For most readers, the attribution question is secondary. The more important point is structural: the aircraft series is volatile because aircraft orders are large and lumpy, regardless of which firm books them. The volatility is a property of the product, not of any one company’s fortunes.
A rule of thumb for the next release
Before reacting to a durable goods headline, pull three lines from the release’s Table 1 and report all three alongside the headline:
- Excluding transportation. This tells you how much of the headline was transportation equipment.
- Nondefense aircraft and parts. This tells you whether the transportation move was aircraft or motor vehicles.
- Nondefense capital goods excluding aircraft. This is the calmer read on business equipment demand.
If the headline and the ex-transportation line diverge sharply, the headline is a transportation story. If the ex-transportation line and the core capital goods line agree, the month’s signal is probably broader than aircraft. If they disagree, look at the other components before writing a sentence about business investment.
The next advance report, covering October 2026, is scheduled for release on November 25, 2026, at 8:30 a.m. EST. The full report with revised and more detailed estimates follows on a later date. The same three lines will be in Table 1.
Why this matters beyond one month
Durable goods orders are one of the more closely watched monthly indicators because they arrive early and they speak to business spending. That combination makes them prone to overreading. A single strong month gets written up as a surge in business investment; a single weak month gets written up as a collapse. The September 2026 release shows why both readings can be wrong: the headline was strong, the ex-transportation line was modest, and the difference was aircraft.
The fix is not to ignore the headline. It is to read it as the top of a decomposition rather than the bottom line. The Census Bureau publishes the components for exactly this reason. The release is a table, not a sentence, and the sentence most outlets quote is the one least suited to telling you what changed.
For the next release, the procedure is short: read the headline, then read the three lines, then decide which story the month actually supports. Most months, the answer will be in the components, not the aggregate.
FAQ
What is the “core capital goods” line in the durable goods report?
It is the M3 aggregate for nondefense capital goods excluding aircraft, published in Table 1 of the advance report. The Census Bureau’s aggregate-series documentation defines the ex-aircraft variant as nondefense capital goods less nondefense aircraft manufacturing. It is a Census Bureau series, not a Federal Reserve or BEA measure of business investment.
Why does the Census Bureau publish an ex-aircraft line at all?
Because aircraft orders are large and lumpy, and the Bureau’s aggregate-series documentation shows the ex-aircraft variant is a defined subtraction of the nondefense aircraft manufacturing series from the nondefense capital goods aggregate. Publishing the ex-aircraft line gives readers a less volatile read on the same underlying survey.
Can I attribute a month’s aircraft swing to a specific company?
Not from the release alone. The M3 release publishes industry aggregates and does not name companies. Any company-level attribution is an inference from outside the release and should be labeled as a possible contributor, not a demonstrated cause.
How precise is the advance estimate?
The M3 survey is not based on a probability sample, so sampling error cannot be measured and confidence intervals cannot be computed. The figures are seasonally adjusted but not adjusted for inflation, and the advance report is revised in the full report published about a week later. Read the advance number as a preliminary signal, not a final measurement.
When is the next release?
The advance report on durable goods for October 2026 is scheduled for November 25, 2026, at 8:30 a.m. EST, per the September 2026 release. The full report with revised and more detailed estimates follows on a later date listed in the release schedule.
Sources
- U.S. Census Bureau, Monthly Advance Report on Durable Goods Manufacturers’ Shipments, Inventories and Orders, September 2026 (CB 26-169, released October 27, 2026), including Table 1 and the explanatory notes on survey design, seasonal adjustment, and release schedule. census.gov/manufacturing/m3/adv/pdf/durgd.pdf
- U.S. Census Bureau, Composition of Aggregate Data Series in Terms of M3 Industry Series, defining NDE (nondefense capital goods) and NXA (nondefense capital goods excluding aircraft) and the industry series that compose them. census.gov/manufacturing/m3/historical_data/aggseries.pdf
- U.S. Census Bureau, Monthly Advance Report on Durable Goods Manufacturers’ Shipments Inventories and Orders release page, October 27, 2026. census.gov/manufacturing/m3/adv/current/index.html