# Can Portland's GDP be divided by business size?

**Finding: no defensible Portland city small-versus-large GDP percentage is available from these sources (high confidence).** [BEA county GDP](https://www.bea.gov/news/2026/gross-domestic-product-county-and-personal-income-county-2024) is by industry and county, not enterprise size or incorporated city. [BEA's small-business page](https://www.bea.gov/data/special-topics/small-business) states the United States lacks a consistent comprehensive small-business measure and regular production of its experimental statistics ceased in February 2026. SUSB's 2022 metro figures—50.2% of employer jobs, 41.0% of employer payroll, 35.4% of employer receipts at firms under 500—answer different questions. Receipts include intermediate inputs, and payroll is only part of value added. Neither can be labeled GDP.

GDP is value added: gross output less intermediate purchases, or compensation plus operating surplus, production taxes less subsidies, and relevant adjustments. It includes government, nonprofits, self-employment and housing services with no clean binary `small` or `large` business assignment. A metropolitan or county GDP denominator also differs from a city policy denominator. A national size ratio multiplied by Portland GDP would compound mismatched size, industry, year, geography and institutional coverage. We **reject** that shortcut, as well as using employment share as value-added share. Historical SBA/BEA national estimates cannot cure the mismatch.

A candidate *research* method would (1) choose county or MSA boundaries, never call them city; (2) obtain establishment/enterprise crosswalk and output or value-added by detailed industry and legal form; (3) estimate industry-specific intermediate-input ratios separately for size classes using confidential Economic Census/BEA or commissioned data; (4) address nonemployer proprietors, nonprofits and multi-establishment allocation; (5) reconcile size-class totals to BEA industry GDP, allocating government and housing to explicit **unclassified** categories; and (6) propagate sampling, disclosure and model uncertainty. SUSB receipts plus national industry input ratios might support a very wide **sensitivity scenario**, but would need validation against BEA totals and tested local input purchases. It is not ready for publication as an estimate.

A less ambitious, publishable alternative is a three-measure panel with exactly named denominators: SUSB employer **jobs**, **payroll**, and **receipts** by enterprise size for a common 2022 metro geography. Display nonemployer county activity separately, never added to the 2022 metro shares. If city administrative aggregates can supply taxable gross receipts by business size and city address, these remain receipts, not GDP; record exemptions and multiple establishments. The first records request in the fieldwork kit seeks a safe aggregate that would improve city measurement without releasing taxpayer identities.

**Decision gate:** Publish a numeric size/GDP split only after obtaining a compatible local size-by-industry value-added source or a transparent, reconciled and sensitivity-tested estimation model. Until then display “unresolved,” the three verified economic shares, and a visual accounting diagram for excluded sectors. This is an information finding, not a claim that small firms' economic role is unimportant.