Learn what monthly US auto sales reports count, how the seasonally adjusted annual rate works and why sales, retail dollars and registrations differ.
What a monthly US auto sales report actually counts
A monthly US auto sales report counts new vehicles delivered to buyers during the month, usually in units, and often restates the total as a seasonally adjusted annual rate. It is a flow, not a snapshot. It describes vehicles delivered to buyers, not the number available to buy.
The source changes the count. A manufacturer reports its own deliveries. A federal agency combines reports into an industry total. A state registration table counts something else entirely.
Units first, then the seasonally adjusted annual rate
The simplest number in a monthly report is the unit count. Toyota, for example, reported 201,306 US sales for September 2026 and gave the change on both a volume and a daily selling rate basis (Toyota release).
The BEA total vehicle sales series publishes an industry figure in millions of units at a seasonally adjusted annual rate. In October 2026 it stood at 16.003 million units. That does not mean 16 million vehicles were sold in October.
A seasonally adjusted annual rate takes a monthly pace, removes the part that follows a recurring seasonal pattern, and expresses the result as if it continued for a year. The Federal Reserve explains that it supplies seasonal factors for car and light-truck sales to the BEA and re-estimates them once a year with X13-ARIMA. The factors are revised, and unusual events can be treated as outliers.
So a SAAR is a statistical summary of one month, not a forecast and not a running annual total.
Why the same month produces two different totals
The BEA publishes an unadjusted series too. In October 2026 the not seasonally adjusted series showed 1,361.913 thousand units. Multiply by 12 and you get about 16.34 million.
The adjusted series showed 16.003 million. The gap comes from seasonal adjustment, not a revision to the count.
| Measure | What it counts | Unit | Example |
|---|---|---|---|
| Manufacturer sales | New vehicles delivered by one brand | Vehicles | 201,306 Toyota and Lexus units, September 2026 |
| BEA total vehicle sales | New vehicles sold, adjusted annual pace | Millions of units, SAAR | 16.003 million, October 2026 |
| BEA unadjusted sales | New vehicles sold, raw monthly total | Thousands of units | 1,361.913 thousand, October 2026 |
| Census retail sales | Revenue at retail stores, by category | Dollars | $745.0 billion total retail and food services, September 2026 |
| FHWA registrations | Vehicles registered by state | Vehicles, annual | 2023 state table |
The Census Advance Monthly Retail Trade Survey measures dollars, not units. Its release says the estimates are adjusted for seasonal variation but not for price changes. A dollar increase can reflect more vehicles, more expensive vehicles, or both.
What manufacturer reports add: selling days and daily selling rate
Selling days vary by month and year. A brand that sells on 25 days has more opportunity than one that sells on 24, even if demand per day is unchanged.
Toyota handles this by reporting two comparisons. Its September 2026 release shows sales up 8.4 percent on a volume basis and up 4.0 percent on a daily selling rate basis. The first uses total units; the second adjusts for selling days.
A volume gain can be smaller or larger once the calendar is taken into account.
Sales, retail dollars and registrations measure different things
A manufacturer’s monthly figure is a delivery count. A Census retail sales figure is revenue. A registration table is a stock of vehicles recorded by state agencies.
The FHWA Highway Statistics table MV-1 lists state motor-vehicle registrations for 2023 and was published in November 2024. It is annual, state-supplied and covers registered vehicles rather than new-vehicle deliveries. The timing alone makes it unsuitable for measuring a single sales month.
Registrations still show where vehicles are registered and how large the fleet is.
Questions to ask before quoting a monthly number
Ask four things. Is the figure units or dollars? Is it adjusted or raw? Does it cover one brand or the whole market? And is it a flow or a stock?
Then check the revision history. Monthly estimates are preliminary and get revised as more responses arrive. The Census release notes that prior-month changes can be revised, and the Fed’s seasonal factors are re-estimated each year.
Finally, separate the number from the interpretation. A higher monthly total does not prove that demand is stronger. It may reflect extra selling days, a model-year clearance, or a comparison with a weak month a year earlier.
For inventory, see what vehicle inventory means for buyers. For price indexes, see how to interpret used car market price indexes. More industry coverage sits in Automotive News and Industry.
Frequently asked questions
Does a monthly sales report count vehicles sold or vehicles registered?
It counts new vehicles delivered during the month, as reported by manufacturers or compiled by a statistical agency. Registrations are a separate, usually annual, count of vehicles recorded by state agencies. The two series answer different questions and should not be substituted for each other.
Why is the seasonally adjusted annual rate different from the monthly total times 12?
Seasonal adjustment changes the monthly figure before it is annualized. The adjustment removes a recurring calendar pattern, so the result differs from a simple multiplication. The BEA publishes both adjusted and unadjusted series so readers can check the difference.
Are monthly auto sales figures final when first published?
No. Early estimates are based on incomplete responses and are revised. Seasonal factors are also re-estimated, which can change adjusted history.
Can a rising sales total mean the market is growing?
Not by itself. A monthly gain can come from more selling days, a weak year-earlier comparison or a model-year clearance. Check the unit count, the adjustment basis and the comparison period before treating a single month as a trend.
