The Bureau of Labor Statistics produces estimates for job openings, hires, and separations by establishment size. These estimates can help to better explain some of the internal dynamics of the labor market. As with the national and regional JOLTS published series, these series start in December 2000. The available files provide users with tables of job openings, hires, and total separations, as well as three breakouts of total separations: quits (voluntary separations), layoffs and discharges (involuntary separations), and other separations. With the JOLTS news release on October 6, 2020, these data have been available in the BLS Labstat database.
If you have any questions on JOLTS establishment size class class data, please email email@example.com.
The production of the JOLTS establishment size class estimates follow standard statistical survey sampling and estimation procedures. (See the JOLTS chapter of the JOLTS chapter of the BLS Handbook of Methods for a more detailed discussion of JOLTS techniques.) These sampling and estimation procedures are explained below.
Sample design–The survey design is a random sample of 21,000 nonfarm business establishments, including factories, offices, and stores, as well as federal, state, and local governments in the 50 states and the District of Columbia. The establishments are drawn from a universe of over 9.1 million establishments compiled as part of the operations of the Quarterly Census of Employment and Wages (QCEW) program. The QCEW program includes all employers subject to state Unemployment Insurance (UI) laws and federal agencies subject to Unemployment Compensation for Federal Employees (UCFE). The scope for these JOLTS size class estimates is limited to establishments in the private sector.
Stratification–The JOLTS sampling frame is stratified by ownership, region, industry sector, and size class. The JOLTS sample is constructed from individual panels of sample units drawn on an annual basis. The full annual sample consists of one certainty panel, composed of only large units selected with certainty based on their size, and 24 noncertainty panels. Each month a new noncertainty panel is rolled into collection, and the oldest noncertainty panel is rolled out. This means that at any given time the JOLTS sample is constructed with panels from up to three different annual sampling frames. Beginning with April 2009, the entire sample is post-stratified and reweighted annually to represent the most recent sampling frame. Additionally, out-of-business establishments are removed from the sample. Since mid-2009, the annual sample is supplemented with a quarterly sample of business birth establishments (i.e., new establishments) so that the JOLTS sample better reflects the actual age distribution of the sample population. NOTE: The sampling weights are assigned at the ownership/region/industry/size class level.
Definition of size class–The maximum employment of the establishment over the last 12 months is used to determine size class at the time of sample selection; this classification stays fixed for a year until the next annual sample is drawn.
Utilizing these size classes, establishments can also be described as small (1-49 employees), medium (50-249), and large (250+).
Data collection–The same data that are collected and used to produce published industry and regional estimates are used for size class estimates. Data are scrutinized for outliers with respect to industry and regional estimates but not for size class estimates.
Aggregated reports–Some sample units provide data at an aggregated level. That is, the respondent provides a consolidated data report. The data for these reports are reweighted accordingly. The same data used for published estimates by industry and region were also used for these size class estimates.
Adjustment for missing data–If there are not sufficient usable responding units in a particular region/industry/size class cell, then data are collapsed across size classes when calculating nonresponse adjustment for missing units. JOLTS imputation methods are applied in the case of item nonresponse.
Benchmarking–The weighted sum of sample-based employment estimates are benchmarked to independent population controls derived from the Current Employment Statistics (CES) survey. The larger CES sample size allows for more reliable population controls. At present size class estimates are not benchmarked at the size class level.
Estimation of levels–The derivation of basic estimates for each respondent is the product of four variables: sampling weight; nonresponse adjustment factor; benchmark factor; and the data element. This product is then summed over all the respondents belonging to each size class.
Birth/Death model component–The Birth/Death model is used to account for contributions from new businesses that cannot be captured by the survey because they are not yet present on the sample frame. The model also adjusts for effects of business closures that may not be captured by the sample as many businesses do not respond to the survey in the month they go out of business. The model estimates are applied at the industry/size class level. At the total private level, the size class estimates of total hires and separations equal those derived from the region/industry level. The sum of the size class estimates of job openings; quits; layoffs and discharges; and other separations, however, differ slightly from those of region/industry as these estimates are based on the distribution of each estimation cell. These estimates are added to the basic estimates for each size class.
Alignment procedure–The JOLTS figure for hires minus separations should be comparable to the CES over-the-month net employment change. Because of its large sample size and annual benchmarking to universe counts of employment from the QCEW program, the CES series is considered a highly accurate measure of net employment change. However, definitional differences, as well as sampling and nonsampling errors between the two surveys, have caused JOLTS to diverge from the CES survey over time. To limit the divergence and to improve the quality of the JOLTS hires and separations series, BLS implemented a monthly alignment method. For size class estimates, this procedure is applied at the total private level such that not seasonally adjusted size class estimates are proportionately aligned to the total private not seasonally adjusted CES industry employment estimates. This procedure is independently applied to each data element.
Seasonal Adjustment–BLS seasonally adjusts the size class series using the X-13-ARIMA seasonal adjustment method. Seasonal adjustment is the process of estimating and removing periodic fluctuations caused by events such as weather, holidays, and the beginning and ending of the school year. Seasonal adjustment makes it easier to observe fundamental changes in the level of the series, particularly those associated with general economic expansions and contractions. A concurrent seasonal adjustment methodology is used in which new seasonal adjustment factors are calculated each month, using all relevant data, up to and including the data for the current month. The six data elements and employment for each size class are independently adjusted.
Estimation of rates–The job openings rate is computed by dividing the job openings level by employment plus job openings, and multiplying this quotient by 100. The hires rate is computed by dividing the hires level by employment, and then multiplying the result by 100. The remaining rates (total separations, quits, layoffs & discharges, and other separations) are computed in the same manner as the hires rate. The employment by size class estimates used in the derivation of the rates were developed from CES employment estimates only for this purpose and are not an official BLS series.
JOLTS estimates are subject to both sampling and nonsampling error as are all sample-based surveys. When a sample rather than the entire population is surveyed, there is a chance that the sample estimates may differ from the "true" population values they represent. The exact difference, or sampling error, varies depending on the particular sample selected. This variability is measured by the standard error of the estimate.
The JOLTS estimates also are affected by nonsampling error. Nonsampling error may occur for many reasons, including the failure to include a segment of the population, the inability to obtain data from all units in the sample, the inability or unwillingness of respondents to provide data on a timely basis, mistakes made by respondents, errors made in the collection or processing of the data, and errors from the employment benchmark data used in estimation.
JOLTS estimates rely on a Balanced-Half sample replication methodology to produce standard error estimates.
Limitations on using the JOLTS size class estimates–Users of these research estimates should be aware of the following:
Last Modified Date: October 14, 2022