Census's new nonemployer data count businesses, not the people working independently

The Census Bureau released its 2024 Nonemployer Statistics on September 29, 2026, providing industry and geographic detail on businesses with no paid employees. It is a useful source for examining economic activity that employer statistics leave out. But the unit being counted is a business tax return, not an individual worker, and the receipts measure should not be mistaken for an owner's take-home income. ^{[1] [2]}

The release covers tax-subject businesses meeting a receipts threshold: at least $1,000 annually, or $1 in construction. It supplies establishment counts and total receipts across 18 industry sectors using the 2022 industry classification. County and metropolitan detail are available, while legal-form breakdowns are limited to national and state figures and receipts-size classes to the national level. These are 2024 business statistics released in 2026, not a contemporaneous employment survey. ^[1]

The distinction begins in the source records. Census builds the nonemployer universe primarily from IRS business-income-tax information rather than conducting a survey to identify these businesses. Its methodology describes screening records to remove employers incorrectly included in that universe. The tabulations have no sampling error because they are not sample-based, but coding mistakes, coverage gaps and misclassification can still affect the results. In particular, Census says it cannot reliably identify every company that leases all its employees. ^[3]

The glossary explains another limit on interpreting the totals as a gig-worker count. NES treats each distinct business income tax return as a firm and uses firm, business and establishment interchangeably. Receipts include gross sales, commissions and business income reported on tax returns. Those definitions describe business records and flows of money; they do not isolate earnings after business costs or identify how many people depend on that work. ^[2]

Location requires similar care. Census says most geographic codes come from a business mailing address, which may differ from the place where operations occur. A county table can consequently describe the recorded business geography without locating every job or customer transaction. The program's overview identifies state and local planning and analysis of independent-worker industries among its uses, but those uses do not remove this limitation. ^{[3] [4]}

Historical comparisons need more than a line drawn through annual totals. Census warns that it does not restate older estimates for changes in industry definitions, geography or methods, and does not recommend treating multiple years as a time series. It updated receipts cutoffs in reference year 2022. Readers comparing an industry with earlier years therefore need to establish whether the relevant definitions and inclusion rules match. ^[3]

The new release points users to tables on data.census.gov, downloadable datasets and an API. When examining small areas or industries, the published flags belong with the figures: the glossary marks withheld or noncomparable values and identifies the degree of confidentiality noise applied to cells. Public access makes the records usable for scrutiny; it does not turn every apparent change into a verified change in workers' livelihoods. ^{[1] [2]}

About this item: Retrospective report. No first-hand reporting or independent testing is claimed. Corrections policy.