LEHD Quarterly Workforce Indicators (QWI) - NAICS 56

The federal record of workforce churn, measured in quarters. The Census Bureau's Longitudinal Employer-Household Dynamics program publishes the Quarterly Workforce Indicators: economic indicators tabulated from administrative records by firm characteristics and worker demographics, covering employment, job creation and earnings alongside the flows underneath them. Every quarter returns employment at the start and end of the quarter, stable full-quarter employment, all hires and new hires, separations, the stable-job turnover rate, firm-level job gains and losses, and average monthly earnings of full-quarter workers - each cut by state, NAICS industry, worker sex, age and education, and firm age and size.

Two properties make the panel hard to substitute. First, the measures are flows observed in the records themselves - a hire and a separation are counted events, not residuals inferred from a level moving between two periods. Second, the breakouts: the same quarter can be sliced by worker trait and by firm age and size, which is where most headline labor series stop.

NAICS 56 - administrative and support and waste management and remediation services - is the staffing, security and janitorial workforce, the highest-churn corner of the commercial-property tenant base and exactly where a level alone tells you least. Twenty-three documented fields ride on every row. Within Datadory's diversified-reits shelf this is the labor-flow fundamental, and it scores 9/10 on our quality rubric.

Each row is one geography x industry x quarter cell. Two real California rows:

geography     : 06   (California)
industry      : 56    # admin & support and waste mgmt & remediation svcs
year          : 1991
quarter       : 4
Emp           : 630048   # jobs on the first day
EmpEnd        : 606362   # jobs on the last day
EmpS          : 457812   # stable full-quarter jobs
HirA          : 300325   # hires, all accessions
Sep           : 324011   # separations, all
EarnS         : 1883     # avg monthly earnings, stable jobs

geography     : 06   (California)
industry      : 00    # all industries combined
year          : 1991
quarter       : 3
EmpEnd        : 12152624
Sep           : 3165558

Read the first row once: 630,048 NAICS 56 jobs at the start of 1991 Q4, 300,325 hires and 324,011 separations inside thirteen weeks, and only 457,812 jobs stable across the whole quarter at an average $1,883 a month. Gross flows equal to roughly half the headcount, in one quarter, in one sector - that is the normal texture of administrative and support services, and it is invisible in any annual or monthly level.

Swap the NAICS code and the same key returns construction, health care or any other two-digit sector; swap the geography and it returns states, counties, metros or workforce-board areas. Name your sector, states and quarter range when you request a sample and Datadory will cut exactly that slice.

Twenty-three documented fields, verified against the released files during the research pass - nothing inferred from column names alone. The identity block keys the cell (periodicity, seasonadj, geo_level, geography, ind_level, industry, ownercode, sex, agegrp, education, year, quarter). The stock block counts what exists (Emp, EmpEnd, EmpS). The flow block measures what changed (HirA, HirN, Sep, TurnOvrS, FrmJbGn, FrmJbLs), and EarnS prices the stable workforce.

The distinction worth internalizing is Emp versus EmpS: point-in-time employment counts every job on day one, while stable full-quarter employment counts only jobs held on both the first and last day with the same employer - and the published turnover rate divides by the stable figure, not the headline.

Finer worker codes (race and ethnicity detail), firm-age and firm-size classes, the net firm-job-change measure and the per-indicator status-flag semantics fold under additional fields on request - their exact shape depends on the cut you specify, so they are confirmed when you ask for a sample.

Geography - all 50 states, DC and Puerto Rico, with county, metropolitan/micropolitan and workforce-board area cuts defined on the same geography codes, so a state view and its metro drill-down stack without remapping.

Temporal - quarterly observations beginning 1990, with the earliest observed files starting in 1991 Q1; start dates vary slightly by state, so the first quarter of your cut gets confirmed at sample request. Current release vintages run R2026Q2/R2026Q3.

Granularity - quarterly cells by state x NAICS industry x worker demographics x firm age and size. This slice holds the two-digit NAICS 56 line; the demographic and firm-trait dimensions are what turn it from a headcount series into a modelable panel, because young-firm hiring and total hiring, or male and female separations, answer different questions.

Set against the wider catalog - average quality score 7.81 across all 1,744 datasets - this slice scores 9/10, carried by complete field documentation and verified sample rows.

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

• Tenant-industry labor-demand monitoring - quarterly employment, hires and separations for the staffing, security and janitorial workforce show whether the service layer under a property type is expanding or quietly thinning.
• Underwriting and credit screens - separation spikes and turnover jumps surface stress in a tenant industry quarters before it reaches an earnings call or a rent roll.
• Labor-availability scoring for site selection - hires by metro and workforce-board area rank local markets on how much staffing-sector hiring actually happens there, not on population.
• Payroll and compensation benchmarking - average monthly earnings of stable full-quarter workers, by market and demographic cut, anchor comp bands for the highest-churn roles in commercial real estate.
• Churn benchmarking and workforce modeling - a stable-job turnover rate computed against stable employment gives planners a like-for-like baseline instead of a naive exits-over-headcount ratio.

Investors and quants model tenant-industry labor momentum at quarterly frequency, with demographic and firm-age cuts the headline series never carries. Data scientists and ML engineers engineer hiring, separation and earnings features off a stable, fully documented schema whose flags travel with every indicator. Developers building data products load cells keyed on geography x NAICS x year x quarter so consecutive releases diff cleanly in the warehouse. Market researchers rank local labor markets by realized hiring in the industries occupying the properties they cover. Journalists, academics and students ground stories on layoffs, hiring waves and wage levels in the official record. Persona workflows sit at investors & quants x diversified REITs, data scientists x diversified REITs, developers & builders x diversified REITs and market researchers x diversified REITs.

Against the rest of the diversified-reits shelf: Business Dynamics Statistics trades frequency for reach - annual firm births, deaths and job flows for the whole economy back to 1978, where this panel answers quarterly and follows the workers. The BLS NAICS 56 employment series reports the monthly level of employment in the sector, not the hiring and separation churn underneath it. ABS Australia - Industry Statistics covers a different continent entirely. Only this panel observes hires, separations and stable-job earnings as they happen, quarter by quarter, with worker demographics attached. The head-to-head with the annual view is worked through in LEHD QWI vs BDS.

Four honest caveats. First, dollars of the period: EarnS is recorded unadjusted for inflation - $1,883 a month in 1991 California is not a modern wage - so deflate before comparing across decades. Second, suppression: cells built on very few jobs carry status flags, and a single-county or single-demographic chart drawn without checking them can be a chart of one employer. Third, uneven starts: the series nominally begins in 1990, but the earliest observed files start 1991 Q1 and first quarters vary by state, so confirm the back-run for your geographies. Fourth, resolution limits: this slice holds NAICS 56 at the two-digit level, and the national aggregate cut is a newer addition to the family - the deep, settled series is state-level. None of these bite unexpectedly; they ship flagged against the analysis you plan to run.

Because the raw artifact ships as roughly 132 files per state per release, keyed by encoded filenames and built for archivists and statistical packages, when most questions want one sector, a handful of states and twenty years of quarters. Datadory cuts samples to your sector, states and quarter range, keys the result for warehouse loading so consecutive releases diff cleanly, and carries the status flags through so nobody charts a one-employer county by accident. Browse the rest of the industry on the diversified REITs data hub, the best diversified-reits datasets ranking, or the full catalog.