The simple version

The Census Bureau publishes monthly orders for durable goods, meaning manufactured items built to last several years. It is one of the earliest reads available on whether businesses are committing to spend.

This morning's report is a clean demonstration of why the headline is not the number to read. Orders for nondefense capital goods, the equipment businesses buy to produce other things, rose 2.0 percent in July. Take aircraft out of that same category and the increase is 0.2 percent.

Nothing was hidden. Census publishes both figures in the same table. One of them travelled into the coverage and the other did not.

The numbers

• New orders for manufactured durable goods increased 1.1 percent in July to $339.3 billion, following a 0.5 percent June increase (U.S. Census Bureau, Advance Report on Durable Goods, July 2026, release CB 26-134, August 26, 2026)
• Excluding transportation, new orders increased 0.4 percent (Census, CB 26-134)
• Transportation equipment led the increase, up 2.3 percent to $116.2 billion, after two consecutive monthly decreases (Census, CB 26-134)
• Nondefense new orders for capital goods increased 2.0 percent to $99.1 billion (Census, CB 26-134)
• The same category excluding aircraft increased 0.2 percent, to $85.9 billion from $85.8 billion (Census, Advance Report Table 1, series labelled Nondefense capital goods, Excluding aircraft)
• New orders for nondefense aircraft and parts increased 12.7 percent in July, and fell 51.1 percent from April to May (Census, Advance Report Table 1)
• The gap between the headline and the same figure excluding transportation is 0.7 percentage points; between the headline and core capital goods orders, 0.9 points; between nondefense capital goods and the same series excluding aircraft, 1.8 points (computed from the Census figures above)
• Census states that statistical significance is not measurable for this survey, because the estimates are not based on a probability sample (Census, CB 26-134)
• Revised and more detailed estimates are published on September 2, 2026, at 10:00 a.m. Eastern (Census, CB 26-134)

Why one category breaks the headline

Most categories in this report behave the way a statistic is supposed to. Orders for machinery, electrical equipment, or fabricated metal move in increments, because many buyers place many orders and the total smooths out.

Aircraft do not work that way. An order is large, the buyers are few, and the timing is driven by contract negotiations and fleet planning rather than by the calendar month. The size of the resulting swings is visible in Census's own table: orders for nondefense aircraft and parts fell 51.1 percent between April and May, then rose 12.7 percent in July.

A category that can halve in a month and then jump by double digits is not describing changing business conditions. It is describing when contracts happened to be signed.

This is a well-known feature rather than a flaw, and Census publishes the components precisely so the effect can be removed. The release gives the headline change and the change excluding transportation in the same sentence, and the table breaks capital goods down further still.

The line that actually gets watched

The series economists follow from this report is the one Census labels nondefense capital goods, excluding aircraft. Each part of that name is removing a specific source of noise.

Capital goods means equipment businesses buy to produce other things, rather than consumer items. Nondefense removes military procurement, which follows budget cycles and policy decisions rather than commercial demand. Excluding aircraft removes the lumpiness described above.

Observers treat what remains as a proxy for what businesses intend to invest in productive capacity. That proxy framing is a convention among people who read the release, not a characterization Census itself makes, and it is worth keeping the distinction. Orders are also commitments rather than completed purchases, so some of them change.

This month is exactly why the distinction earns its keep. Read the capital goods headline and business investment looks like it accelerated. Read the line beneath it and the same category was close to flat, with aircraft accounting for almost the entire difference.

The corporate profits figure inside this morning's gross domestic product revision, which we cover separately, has the same shape. In both cases the release contains a sub-series that answers the question better than the number on top of it, and in both cases the sub-series gets less attention.

The Real Cost lens on a business-investment signal

Business equipment orders are not a household number, and the connection is indirect enough to state carefully rather than overstate.

• Orders for business equipment are a commitment to spend, which typically precedes hiring and production at the firms filling those orders
• That makes the series an early signal about industrial employment, arriving before the payroll data shows anything
• It is a signal rather than a forecast, and orders can be revised, delayed, or cancelled between commitment and delivery
• For a household, the practical use is context: a flat core reading under a healthy headline usually means the headline is describing aircraft rather than the economy

That last line is the whole article compressed. When this report produces a number worth reacting to, the first move is to check whether transportation explains it, and the release answers that in the same table.

What this means

When a durable goods headline lands, read two numbers instead of one: the headline, and the change excluding transportation. Census prints both, and the gap between them tells you how much of the story is aircraft. This month that gap was 0.7 percentage points at the top line and 1.8 points inside capital goods.

The broader habit is that most government releases contain the correction for their own biggest distortion, published in the same document. The distortion travels into headlines, and the correction sits one table down, free, unread.

What this is NOT

This is not a prediction of business investment, manufacturing activity, employment, or the economy. This is not a claim that the Census Bureau publishes anything misleadingly; the agency publishes every component, and the excluding-transportation figure appears in the same sentence as the headline. This is not advice about any security, fund, or financial decision, and no manufacturer or company is named or evaluated. This is not a position on manufacturing, trade, or industrial policy. Orders are commitments rather than completed transactions and are subject to revision and cancellation. The proxy reading of the core series is a convention among observers rather than a characterization by Census. Figures are as published on the dates stated. This is not investment or financial advice of any kind.

Sources

• U.S. Census Bureau, Advance Report on Durable Goods Manufacturers' Shipments, Inventories and Orders, July 2026 (CB 26-134), released August 26, 2026: https://www.census.gov/manufacturing/m3/adv/pdf/durgd.pdf
• U.S. Census Bureau, Advance Report current release and data tables: https://www.census.gov/manufacturing/m3/adv/current/index.html
• U.S. Census Bureau, Manufacturers' Shipments, Inventories, and Orders (M3): https://www.census.gov/manufacturing/m3/index.html
• U.S. Census Bureau, M3 how the data are collected: https://www.census.gov/manufacturing/m3/how_the_data_are_collected/index.html