B. BEA Industry Accounts Reference
This appendix provides reference tables for interpreting BEA industry data, which underpins much of the analysis in this book.
GDP by Industry (2023)
The following table shows value added (GDP contribution) by major industry, sorted by size.
Real estate, rental, leasing
$2,843
10.4%
+25%
53
Professional, scientific, technical services
$2,426
8.9%
+35%
54
Government
$2,217
8.1%
+10%
—
Finance and insurance
$2,191
8.0%
+22%
52
Health care and social assistance
$2,074
7.6%
+18%
62
Manufacturing
$2,041
7.5%
+12%
31-33
Retail trade
$1,573
5.8%
+20%
44-45
Information
$1,471
5.4%
+45%
51
Wholesale trade
$1,433
5.2%
+15%
42
Construction
$1,184
4.3%
+28%
23
Transportation and warehousing
$926
3.4%
+22%
48-49
Other services
$640
2.3%
+8%
81
Accommodation and food services
$622
2.3%
+5%
72
Administrative and waste services
$552
2.0%
+18%
56
Utilities
$391
1.4%
+6%
22
Management of companies
$384
1.4%
+30%
55
Educational services
$375
1.4%
+12%
61
Arts, entertainment, recreation
$303
1.1%
+15%
71
Mining
$290
1.1%
-20%
21
Agriculture, forestry, fishing
$243
0.9%
+5%
11
Source: BEA Industry Accounts, 2023 data. Growth rates are approximate 10-year real (chained 2017 dollar) growth, 2013-2023.
Note: Information (NAICS 51) shows the fastest growth, driven by software, data processing, and streaming. Mining's decline reflects the oil price collapse of 2014-2016 and structural shift away from coal. Accommodation and food services shows weak growth because of the COVID-19 shock, from which the sector had not fully recovered in value-added terms by 2023.
Manufacturing Subsectors
Chemical products
$382
325
Computer and electronic products
$291
334
Food, beverage, tobacco
$251
311-312
Motor vehicles and parts
$195
336
Petroleum and coal products
$176
324
Machinery
$153
333
Miscellaneous manufacturing
$133
339
Fabricated metal products
$127
332
Plastics and rubber products
$87
326
Electrical equipment, appliances
$74
335
Primary metals
$59
331
Nonmetallic mineral products
$58
327
Paper products
$52
322
Printing and support activities
$33
323
Furniture and related products
$30
337
Textile mills and products
$21
313-314
Wood products
$38
321
Apparel and leather
$11
315-316
Source: BEA Industry Accounts, 2023 data
State GDP Rankings (2023)
1
California
$4,080
14.6%
2
Texas
$2,356
8.4%
3
New York
$2,140
7.7%
4
Florida
$1,523
5.5%
5
Illinois
$1,050
3.8%
6
Pennsylvania
$931
3.3%
7
Ohio
$800
2.9%
8
Georgia
$788
2.8%
9
New Jersey
$752
2.7%
10
Washington
$741
2.7%
Source: BEA Regional Accounts, 2023 data
Metropolitan Area GDP (Top 20, 2023)
1
New York-Newark-Jersey City
$2,160
+8%
2
Los Angeles-Long Beach-Anaheim
$1,105
+6%
3
Chicago-Naperville-Elgin
$886
+5%
4
Washington-Arlington-Alexandria
$715
+7%
5
San Francisco-Oakland-Berkeley
$696
+4%
6
Boston-Cambridge-Newton
$641
+10%
7
Dallas-Fort Worth-Arlington
$613
+15%
8
Philadelphia-Camden-Wilmington
$558
+5%
9
Houston-The Woodlands-Sugar Land
$558
+9%
10
San Jose-Sunnyvale-Santa Clara
$467
+7%
11
Seattle-Tacoma-Bellevue
$450
+11%
12
Atlanta-Sandy Springs-Alpharetta
$449
+14%
13
Miami-Fort Lauderdale-Pompano Beach
$415
+16%
14
Phoenix-Mesa-Chandler
$310
+18%
15
San Diego-Chula Vista-Carlsbad
$283
+9%
16
Detroit-Warren-Dearborn
$280
+4%
17
Minneapolis-St. Paul-Bloomington
$278
+6%
18
Denver-Aurora-Lakewood
$255
+10%
19
Baltimore-Columbia-Towson
$225
+3%
20
Tampa-St. Petersburg-Clearwater
$204
+19%
Source: BEA Metropolitan Area Accounts, 2023 data. Growth rates are approximate 5-year real GDP growth (2018-2023).
How to Read BEA Industry Tables
Value Added vs. Gross Output
Value added (GDP): The contribution to final output. Value added = Gross output minus intermediate inputs. This is what appears in the GDP tables.
Gross output: Total production, including intermediate goods. Gross output counts all sales, including business-to-business transactions.
Example: A steel mill sells $100 million in steel to an auto plant. The auto plant uses that steel to make cars worth $200 million.
Gross output: $300 million ($100M steel + $200M cars)
Value added: $200 million (only the final product counts)
Real vs. Nominal
Nominal (current dollars): Values at prices of the measurement year
Real (constant/chained dollars): Values adjusted for inflation, typically using 2017 as base year
Use real values for growth comparisons over time; nominal for current-period analysis.
Industry Classification
BEA uses NAICS (North American Industry Classification System). Key features:
Two-digit codes: Broad sectors (e.g., 31-33 = Manufacturing)
Three-digit codes: Subsectors (e.g., 334 = Computer and electronic products)
Four-digit codes: Industry groups
Five/six-digit codes: Detailed industries
NAICS is revised every five years. Historical comparisons may require concordance tables when industry definitions change.
Input-Output Tables
BEA produces input-output tables showing inter-industry purchases. Two main types:
Use Table: Shows which industries purchase which commodities
Rows: Commodities (products)
Columns: Industries (producers)
Make Table: Shows which industries produce which commodities
Rows: Industries
Columns: Commodities
The symmetric industry-by-industry table (derived from Use and Make tables) shows direct purchases between industries.
Reading the I-O Tables
A cell showing $50 billion at the intersection of "Automotive" row and "Steel" column means the automotive industry purchased $50 billion worth of steel products.
Multipliers
BEA calculates economic multipliers showing total impact (direct + indirect + induced) of changes in final demand. Type I multipliers capture direct and indirect effects; Type II adds induced effects from household spending.
Measurement Challenges
Several industries pose particular measurement difficulties that readers should understand when interpreting GDP-by-industry figures:
Financial Services (FISIM): BEA measures financial intermediation output using Financial Intermediation Services Indirectly Measured (FISIM)---the difference between interest rates charged on loans and paid on deposits. This is an imputation, not a direct transaction. When interest rate spreads narrow, measured financial output falls even if banks process the same volume of transactions. FISIM can swing substantially with monetary policy changes.
Real Estate (Imputed Rent): The largest single component of real estate GDP is imputed rent---the rent homeowners are deemed to pay themselves. This accounts for over $2 trillion in GDP and exists to make GDP comparable across countries with different homeownership rates. No money changes hands; BEA estimates what owner-occupied homes would rent for based on comparable rental units.
Government: Government output is measured at cost (compensation + depreciation), not at market value, because most government services are not sold. This means government "productivity" cannot increase by definition---a faster-working bureaucrat produces the same measured output as a slower one. This convention understates the contribution of government when public services improve and overstates it when they deteriorate.
Information Technology (Hedonic Adjustment): BEA uses hedonic price indices to adjust for quality improvements in computing equipment. A laptop that costs $1,000 today is vastly more powerful than a $1,000 laptop from 2010. Hedonic adjustment attempts to measure this quality gain, which can cause real IT output to grow much faster than nominal spending suggests. The resulting figures can appear counterintuitive: real computer output may grow 15-20% annually even as spending grows modestly.
Healthcare: BEA measures healthcare GDP as the value added by healthcare providers (hospitals, physician practices, pharmacies). This differs from the commonly cited "healthcare is 18% of GDP," which refers to CMS National Health Expenditure, a broader measure that includes insurance administration, out-of-pocket spending, and government transfers. See Chapter 1 for further discussion.
Data Vintage Notes
BEA revises data through a predictable schedule:
GDP: Three monthly estimates (advance, second, third), then annual and comprehensive revisions
Industry accounts: Annual revisions, quinquennial benchmark revisions
Regional accounts: Quarterly for states, annual for metros
Always note the vintage of data used in analysis. Current estimates may differ from figures cited in this book.
For current data, visit bea.gov
Last updated