U.S. Census Annual Retail Trade Survey / AIES - Motor Vehicle Dealers
Datadory delivers automotive retail data covering the U.S. Census Bureau's annual benchmark of motor vehicle and parts dealer firms: sales, e-commerce sales, gross margin, purchases, operating expenses and year-end inventories for NAICS 441 down to six-digit lines such as new car dealers (44111) and used car dealers (44112), from 1992 onward - request a sample from Datadory.
What this dataset is
The U.S. Census Annual Retail Trade Survey / AIES - Motor Vehicle Dealers record is the annual benchmark behind every other US auto-retail number you will meet. From 1992 through reference year 2022 it ran as the Annual Retail Trade Survey (ARTS); it has since transitioned into the Annual Integrated Economic Survey (AIES), whose first main release - reference year 2023 - arrived on 26 February 2026. The survey covers roughly 16,500 employer firms sorted into 85 industry strata, and its estimates are what the monthly retail surveys are reconciled against.
For motor vehicle and parts dealers the kind-of-business detail runs from the two-digit NAICS 441 headline down to 4411 automobile dealers, 44111 new car dealers, 44112 used car dealers, 4412 other motor vehicle dealers (including 44121 recreational vehicle dealers) and 4413 automotive parts, accessories and tire stores. Measures ride along together: sales, e-commerce sales, purchases, gross margin and gross margin as a percent of sales, per capita sales, sales taxes, year-end inventories, payroll and employment, plus total operating expenses with the detailed breakout published in years ending in 2 and 7.
Datadory scores this record 9 out of 10 - verified field definitions, deep history, annual cadence. Request a sample and start with the field dictionary below.
Sample rows
Five verified rows ship with the record. The top block is the AIES reference-year 2023 print at three NAICS levels; values are thousands of dollars except employees. The bottom pair comes from the ARTS restated sales tables, where each measure workbook carries one row per kind-of-business line and one column per year:
NAICS kind of business year sales $1k gross margin ecommerce inventories purchases employees
-----------------------------------------------------------------------------------------------------------------------------
441 Motor vehicle and parts dealers 2023 1,608,512,454 496,349,920 68,664,141 218,465,144 1,159,487,682 2,011,481
44111 New car dealers 2023 1,217,211,141 - - - - 1,119,109
44112 Used car dealers 2023 160,351,361 - - - - 164,333
ARTS restated tables (sales only, millions of dollars)
-----------------------------------------------------------------------------------------------------------------------------
441 Motor vehicle and parts dealers 2022 1,495,937
4411 Automobile dealers 2022 1,273,740Read across the first row: $1,608.5 billion in motor vehicle and parts dealer receipts against $496.3 billion of gross margin (30.9 percent), $68.7 billion of that transacted online, $218.5 billion sitting on lots and lots at year end, and just over 2.0 million employees on the mid-March payroll. Split the headline and new car dealers take $1,217.2 billion of receipts with 1.12 million workers, while used car dealers book $160.4 billion with 164,000. Every figure above reproduces exactly in the sample Datadory returns.
What fields does this dataset include?
Ten fields are verified against the published record layout; they identify the row, date it and carry the estimate. The remaining measures arrive as additional columns described under 'additional fields on request', so nothing in your contract depends on a column we have not shown you.
How far does the coverage reach?
Geography: United States totals by kind of business, with selected sub-national geographies available on request for the AIES era.
Temporal: ARTS-era series run long - sales, inventories and purchases from 1992, gross margin from 1993, per capita sales from 2000, e-commerce from 1998 and sales taxes from 2004, all to 2022. AIES continues the panel from reference year 2023 forward. Detailed operating expenses appear only in years ending in 2 and 7, so expense modeling works on a five-year lattice.
Granularity: annual estimates at seven kind-of-business levels (441, 4411, 44111, 44112, 441120, 44121, 4413), each carrying the full measure set where defined. One sample request covers the whole grid; there is no establishment or dealership layer here, only firm-class aggregates.
How do you get this data?
API, files, or your warehouse. Daily, weekly, or hourly.
Who uses this data?
Eight Datadory personas qualify this record in automotive retail, ranked by assigned relevance:
- Market Researchers & Consultants (relevance 3) - size the US dealer economy by line of business with margin structure attached, then defend the numbers with a federal citation. Our market researchers use cases page shows the workflow.
- Investors & Quant Researchers (relevance 3) - calibrate annual auto-sector models on the same series federal statisticians use to revise the monthly prints; see investors quants use cases.
- Data Scientists & ML Engineers (relevance 3) - join thirty years of sales, e-commerce share and inventory levels into one annual panel for structural demand models; continue at data scientists use cases.
- Journalists, Academics & Students (relevance 3) - cite employer-firm benchmarks that carry published sampling-reliability figures rather than trade-association estimates.
- Competitive Intelligence & Product Teams (relevance 2) - separate category growth from company performance in quarterly reviews; method notes live at competitive intel product teams use cases.
- Developers & Data-Product Builders (relevance 2) - wire documented, stable column names into pipelines without re-deriving definitions each release; see developers builders use cases.
- Sales & Growth Teams (relevance 2) - target territories by dealer-line revenue and payroll instead of registration counts alone.
- E-commerce Operators (relevance 2) - track the parts and accessories e-commerce line to time aftermarket assortment bets against a measured online-share curve.
How does it compare to alternatives?
Inside the Census trio the split is cadence versus depth versus geography. The Monthly Retail Trade Survey record moves monthly but stops at sales, inventories and their ratio; County Business Patterns (also scored 9) adds establishment, employment and payroll counts by county without any dollar sales. This record is the only one of the three that reaches gross margin, purchases, operating expenses and e-commerce receipts - which is why it anchors annual sizing work while the others carry frequency and geography. We score the full trade-off in our vs U.S. Census Monthly Retail Trade Survey - Motor Vehicle and Parts Dealers comparison.
Trade-association alternatives go deeper into the dealership P&L but narrower in coverage: NADA Data's annual financial profile reports department-level detail for a member-defined dealer population, while this record covers roughly 16,500 employer firms with published error measures. For non-US markets, the UK SMMT registrations hub answers fleet-turnover questions rather than revenue ones.
What should you know before relying on it?
Three caveats, stated plainly. First, the program changed skin: ARTS ran to reference year 2022 and AIES takes over from 2023, so measure-by-measure histories splice across the two - keep the series definitions from the field dictionary when you bridge them.
Second, some measures are lumpy by design. Detailed operating expenses publish only in years ending in 2 and 7, and whether AIES will sustain full six-digit NAICS detail beyond 2023 is not yet documented - treat the finest lines as confirmed through 2023 rather than promised forward.
Third, these are survey estimates of employer firms: values in current dollars, sampling error quantified on the sales line via a published coefficient of variation, and no establishment or dealership-level records anywhere because respondent returns stay confidential. None of this dents why the record earned its 9/10: verified field definitions, five sample rows visible before you commit, and coverage stated to the year.
Questions buyers ask
What exactly does this dataset measure?
Annual measures for US motor vehicle and parts dealer employer firms: sales, e-commerce sales, purchases, gross margin, operating expenses, per capita sales, sales taxes, year-end inventories, payroll and employment, published at NAICS 441 down to six-digit lines such as 44111 new car dealers.
How far back does the history go?
Sales, inventories and purchases run from 1992, gross margin from 1993, per capita sales from 2000, e-commerce from 1998 and sales taxes from 2004, through the 2022 ARTS close-out; the Annual Integrated Economic Survey continues the panel from reference year 2023 forward.
Does it include new and used car dealers separately?
Yes. The 441 headline splits into 4411 automobile dealers, then 44111 new car dealers and 44112 used car dealers, alongside 4412 other motor vehicle dealers (with 44121 recreational vehicle dealers) and 4413 automotive parts, accessories and tire stores.
Is dealer-level or store-level data included?
No. The floor is the kind-of-business aggregate: respondent returns stay confidential, so the record contains firm-class estimates for roughly 16,500 employer firms rather than identifiable dealerships. Sampling reliability for the sales estimate ships as a published coefficient of variation.
How does it differ from the monthly Census auto dealer data?
Cadence and depth. The monthly record delivers sales, inventories and inventory-to-sales ratios twelve times a year; this record lands once annually but adds gross margin, purchases, operating expenses, e-commerce receipts and taxes - and serves as the benchmark the monthly estimates reconcile against.
Can I get a sample before committing?
That is the point of this page. Request a sample and Datadory returns real rows - the field dictionary above shows which columns arrive, and the sample table shows the values you should expect to reproduce, from the 2023 AIES print back through the restated ARTS tables.
See the rows before you pay anything.
Name this dataset and we send real records from it — scoped to the fields you asked for.