US BEA — GDP & Industry Accounts for Printing

Most print-industry data counts printers - establishments, presses, shipments. This record answers a different question: what printing is worth inside the world's largest economy. It is the Bureau of Economic Analysis's GDP-by-industry account cut to NAICS 323, showing each industry's contribution to GDP as value added alongside employee compensation, gross operating surplus, taxes on production, gross output and intermediate inputs, in current and chained (real) dollars, with KLEMS production statistics and employment measures in the same dataset family.

Two details make the printing cut usable rather than decorative. First, industry detail follows BEA's own codes with a published concordance onto NAICS, so printing and related support activities (NAICS 323) isolate as a clean line instead of hiding inside a manufacturing aggregate. Second, the accounts report every US industry on identical definitions, so printing's gross output sits next to packaging's, media's and logistics' in the same units and vintages - which is what makes cross-industry denominators computable instead of improvised.

Inside a catalog of 1,744 datasets across 159 viable industries, this record scores 9 out of 10 on Datadory's rubric with verified field definitions - a standard only 85.7% of the catalog meets. Dozens of tables span the headline, underlying-detail and input-output dataset families, and a single full-table pull returns thousands of cells. See where it sits in the commercial printing data hub, then get a sample of this dataset.

One row per industry-table-period observation, flat and immediately pivotable. Illustrative rows in the delivered column order - the ten-field dictionary beneath them is verified against the publisher's documentation, and the values themselves wait for your extract:

TableID    : <GDP-by-industry table>     Frequency   : A | Q
Year       : <years you name>            Industry    : <BEA code on the NAICS 323 line>
TimePeriod : YYYY | YYYYQn               Metric_Name : current dollars | chained (real) dollars
CL_UNIT    : <calculation basis>         UNIT_MULT   : 6 means millions
DataValue  : confirmed with your sample  NoteRef     : <pointer into the response notes>

Three things worth noticing. First, the accounting identity underneath: gross output minus intermediate inputs lands at value added, which then splits into employee compensation, gross operating surplus and taxes on production - so labor share, margin and tax burden arrive as subtractions and groupings rather than separate purchases. Second, the dollar dual: levels run in both current and chained (real) dollars, which separates price effects from volume effects without a deflation project of your own. Third, UNIT_MULT - a base-10 exponent carried beside every figure, so 6 marks millions and no magnitude ever depends on guessing the publisher's scaling convention.

Our catalog ships no preview rows for this record, and pretending otherwise would be worse than useless: live values for the tables, industries and years you name travel with the sample, pinned against the extract before anything downstream depends on them. The structure traces to the August 2026 verification pass.

Ten documented fields define every observation, and they divide cleanly by job: four address the cell (TableID, Frequency, Year, Industry), five describe the measurement (DataValue, TimePeriod, Metric_Name, CL_UNIT, UNIT_MULT), and NoteRef points into the notes that carry table titles. Field definitions on this record carry a verified flag in our catalog - checked against the publisher's own documentation rather than inferred from outputs, which only 85.7% of the 1,744 cataloged datasets earn.

Because each observation is one industry-table-period member, the corpus pivots without reshaping: industries down the side, years across the top, whichever table's measure you want as the values. Deeper attributes fold under additional fields on request - the finer underlying-detail industry tables, which BEA itself cautions run at lower quality than its published aggregates, and the input-output supply-use views can ride along in the same extract.

• Geography: the United States as a single national economy - industry accounts, not a regional grid, so state and county resolution lives in a separate dataset family rather than inside this one. For printer-counted geography, pair it with Census Economic Census, ASM & County Business Patterns.
• Temporal: annual estimates run 1997 through the present, with the percent-change and contributions tables starting in 1998; quarterly estimates reach back to 2005. Nearly three decades of annual structure plus two decades of quarterly turns for the same six-digit industry is depth most print sources cannot approach.
• Granularity: industry x table x year-or-quarter. Industries key on BEA's own codes with a published concordance onto NAICS, which is what isolates printing and related support activities as a clean line instead of a manufacturing residual.
• Scale: dozens of tables across the GDPbyIndustry, underlying-detail and input-output dataset families; a single full-table pull returns thousands of cells.

API, files, or your warehouse. Daily, weekly, or hourly.

Cadence is your call regardless of the rhythm upstream: these accounts describe structure, and your models should get them on whatever schedule your alerts and pipelines expect, shaped as CSV, Parquet or JSON. Name the tables, industries and years when you request the sample - value added for NAICS 323 across the full annual run, quarterly gross output since 2005, or the entire grid - and it ships cut to that universe in exactly the ten-field shape above. The sample comes first either way; the recurring feed follows the same shape. Get a sample of this dataset.

• Market sizing with a federal anchor - value added and gross output give "how big is US printing" an economy-accounting answer instead of a stitched trade estimate, and the same tables size every competitor industry on identical definitions; methods continue on our market sizing page.
• Demand-side factors for models - quarterly real gross output for NAICS 323 works as an exogenous feature behind advertising, media and packaging exposures, the classic move for investors and quants (quant backtesting).
• Margin and labor-share analysis - the value-added decomposition splits printing's dollar into employee compensation, gross operating surplus and taxes on production, so profit pool and labor intensity read straight off the accounts rather than out of a survey sample.
• Cross-industry benchmarking - because every US industry reports on the same definitions, printing's share of GDP, of gross output or of intermediate-input spending is a grouping exercise, not a harmonization project.
• Citation-grade commentary - "printing's shrinking share of US GDP" is a claim this record substantiates with an institution standing behind every figure; see the pattern in citation-grade research.

Ranked by fit in Datadory's tagging for this slice: market researchers and consultants (relevance 3) get the federal anchor that turns a print market model from estimate into account - see market researchers use cases. Developers and data-product builders (relevance 3) stand economics dashboards on a documented table-and-field structure whose layout repeats vintage after vintage - see developers builders use cases. Journalists, academics and students (relevance 3) cite national-accounts figures with the publishing institution named (journalists academics use cases). Data scientists and ML engineers (relevance 2) inherit typed, unit-explicit series that drop into panels without a cleaning chore (data scientists use cases). Investors and quant researchers (relevance 2) treat quarterly real output as a demand-side factor with decades of history (investors quants use cases). Competitive-intelligence and product teams (relevance 1) test category volume assumptions against the industry's aggregate real-output trend before committing to expansion plans.

Methodology note - know what this record measures: economic structure, not turning points. National accounts describe how large printing is and how its dollar divides, at a periodicity coarser than monthly price and employment series; for month-by-month movement, the BLS PPI, CES & OES record is the complement, and the head-to-head sits on our comparison page.

Completeness note - two items were left open at the August 2026 verification pass and are settled at sampling rather than papered over here: the catalog ships no preview rows for this record, and whether printing stands as its own line in the newest vintages of every table - versus rolling up inside a manufacturing aggregate in some - gets confirmed against live records when your sample is prepared. The publisher's own caution about its finer underlying-detail tables travels with any extract that includes them.

Where to go next - pair the value-added layer with the counted baseline: US Census Bureau Economic Census, ASM & County Business Patterns (NAICS 3231) for establishments and payroll, and the source story on the BEA profile. The ranked slate sits on best commercial printing datasets, the long-form version lives in the commercial printing data guide, and the rail below collects the neighbors.