--- title: "Industry data — how county industry figures are built | Data Tactical Group" route: "/docs/methodology/industry" updated: "2026-09-04" --- # Industry data What DTG's industry measures count, and how the published record is completed. Updated 2026-09-04 DTG industry data reports employment, establishments and earnings by industry for every county and county-equivalent, every state, and the nation, quarterly and annually, from 2004 forward. This page describes what each measure counts, how industries and geographies are classified, and how DTG completes the parts of the published record that the Bureau of Labor Statistics withholds. The source of record is the Quarterly Census of Employment and Wages. The [data sources](/docs/methodology/sources) page describes it and every other input. ## Measures **Establishments.** BLS defines an establishment as a single economic unit — a farm, a mine, a factory, a store — that produces goods or services. Establishments are typically at one physical location and engaged in one, or predominantly one, type of economic activity. An establishment is distinct from a firm or company, which may consist of several establishments. QCEW publishes establishment counts for every industry and area it reports, and DTG reports them as published. **Employment.** QCEW reports employment for each of the three months in a quarter. A DTG quarterly employment figure is the average of those three months. Employment counts jobs covered by state unemployment insurance programs, not workers: a person holding two covered jobs is counted twice. **Total quarterly wages.** All wages paid during the quarter by establishments in the industry and area, as reported for unemployment insurance purposes. Employee contributions to Social Security, health insurance, unemployment insurance, workers’ compensation and private pension and welfare funds are included, as are income taxes and union dues withheld from gross pay. Employer contributions for those same purposes are not reported as wages. **Average pay.** Total quarterly wages divided by average quarterly employment. This is a per-job figure for the quarter, not an annual salary and not an hourly rate. **Annual figures.** Annual employment is the average of the four quarterly employment figures. Annual wages are the sum of the four quarters. Annual average pay is the sum of four quarters of wages divided by the average of the four quarterly employment figures. DTG data is not seasonally adjusted. A quarterly figure reflects the actual level in that quarter, including seasonal movement. Comparing a quarter to the same quarter a year earlier is usually more informative than comparing it to the quarter before. ## Industry classification Industries are classified using the North American Industry Classification System (NAICS). DTG carries every level QCEW publishes, from 2-digit sectors through 6-digit national industries, at every geography including counties, as well as the total, domain and supersector aggregations. NAICS 99, covering establishments that could not be classified into an industry, is carried as its own code. It is included in totals. A reader summing detailed industries to a sector total should expect to find it there. ## Geographic coverage Industry data is reported for every county and county-equivalent in the fifty states and the District of Columbia. County-equivalents include parishes, boroughs and census areas, independent cities, and Connecticut’s planning regions. Each state and the District of Columbia, except Utah, also carries a single **unallocated area**, holding employment and wages reported to that state but not attributable to a specific county. It is an ordinary area for data purposes and is included in state and national totals, but it has no geographic extent: it appears on no map and has no public data page of its own. A reader summing counties to a state total should include the unallocated area where the state has one. Counties alone will fall short of the state. ## Geographic and classification consistency Within a DTG vintage, every year is expressed in a single classification. The NAICS and county definitions used are those in effect for the most recent QCEW release in that vintage, and the entire series from 2004 forward is restated onto them. A figure for 2007 is therefore reported using today’s industry codes and today’s county boundaries, not the codes and boundaries in effect in 2007. This makes any two years in a vintage directly comparable without the reader having to reconcile a classification change. It also means a DTG figure for an earlier year may differ from the figure BLS published for that year at the time, where the classification change touched it. Values affected by restatement carry flags — one for industry reclassification, one for county-definition changes. Each flag is applied individually, and only where the value itself changed: if a detailed industry was restated but its parent total was not affected, the parent is not flagged. The flags are available through the API and in exports, and are described further on the [disclosure and provenance](/docs/methodology/disclosure) page. When BLS adopts a new NAICS revision or a county definition changes, the next DTG vintage restates the full series accordingly. ## Suppression QCEW withholds a value where publishing it could reveal information about an individual employer. Suppression is a legal requirement on BLS, not an editorial choice, and it is common. Its incidence rises sharply as areas get smaller. Characteristically, around 81 percent of industry values are withheld and estimated in counties with fewer than 1,000 covered workers, against around 35 percent in counties with 500,000 or more. The exact shares differ from one vintage to the next; the pattern does not. Suppression is heaviest precisely where industry-specific analysis has the fewest alternatives. QCEW does not suppress establishment counts. An industry in a county may report its establishments while the employment and wages associated with them are withheld. ## Estimation DTG estimates the values QCEW withholds rather than leaving them empty, and labels every estimated value as estimated. Published QCEW values are never overwritten. Estimation applies only where QCEW does not disclose a value. Estimates draw on five inputs: - the same industry and area in earlier and later published periods - disclosed values at adjacent levels of the NAICS hierarchy - disclosed values in neighboring and containing geographies - QCEW establishment counts, which are published even where employment is not - County Business Patterns These inputs produce initial estimates, which are then fitted so that the completed data satisfies every constraint the published record imposes. The industry hierarchy and the geographic hierarchy are constrained simultaneously, not one after the other: detailed industries sum to their published parent, counties and the unallocated area sum to their state, and states sum to the nation, at every level of both dimensions at once. The same constraints apply to employment and to wages. Employment sums to within rounding; earnings and establishment counts sum very closely but not always exactly, and the differences are largest in the smallest industries. Estimates are non-negative and are bounded by the range the disclosed values around them permit. Where a published parent total is known, the estimated values beneath it cannot exceed what that total leaves available. The fitting process builds on established statistical techniques for reconciling incomplete tabular data against known totals, extended with methods developed at DTG. The implementation is proprietary and is not published. ## Reading estimated values Every value carries a disclosure code identifying where it came from, and every estimated value in a published QCEW quarter carries a reliability score indicating how tightly the surrounding published data constrains it. Both are described on the [disclosure and provenance](/docs/methodology/disclosure) page. Two points are specific to industry data. A zero in a disclosed value is a measured zero: QCEW reported no covered employment for that industry in that county. A very small estimated value means the estimate is small, not that the industry is absent. Because QCEW publishes establishment counts regardless of suppression, those counts are a useful check. Establishments present alongside a small or estimated employment value indicate real but limited activity. No establishments indicate the industry is genuinely absent from the county in that quarter. Still stuck? A person can help. [Get support](/support)