The JOLTS program at the Bureau of Labor Statistics is currently pursuing research on the number of job openings, hires, and separations by firm size. These estimates may help to better explain some of the internal dynamics of the labor market. These experimental series begin with December 2000 (also the starting point of the official JOLTS series) and end with December 2016. The available files provide users with tables and charts 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. This document presents and briefly discusses the firm size estimates then details the methodology used to create the estimates.
JOLTS has designated these data by firm size as experimental because the research to date has been completed in an environment outside the normal monthly production process. Therefore, these data are not considered official BLS series. For more information on the official JOLTS series and more detailed descriptions of the JOLTS data elements, see the JOLTS chapter of the BLS Handbook of Methods.
JOLTS data users interested in these experimental firm size class estimates can download excel files of seasonally adjusted estimates and not seasonally adjusted estimates. BLS welcomes user comments and inquiries on these data, the methodologies, and the utility of the data via e-mail to firstname.lastname@example.org or phone call to (202) 691-5870.
Size of firm has the largest effect on the JOLTS job openings data element. Firm size 3 (500+ employees) has over twice as many job openings as firm size 1 (1–49 employees) and firm size 2 (50–499 employees). Firm size 2 posts slightly more job openings each month than firm size 1. Job openings declined during the most-recent recession in all three firm sizes, but firm size 3 cut job openings at a faster pace. Similarly, all three firm sizes increased job openings after the end of the most-recent recession, but the firm size 3 had the strongest recovery. Looking at job openings rates, the firm size 3 has the highest job openings rate, followed by firm size 2, then firm size 1 for the entire series.
The number of hires by firm size follows the same patterns as job openings. Firm size 3 has the largest number of hires per month—nearly twice as many as firm size 1 and firm size 2. Hires for firm size 3 declined more during the most-recent recession but also rose more post-recession. The hires rates, however, are nearly equal from 2013 onward for all firm sizes.
Total separations show similar patterns as job openings and hires. Firm size 3 has nearly twice as many separations each month as firm size 1 and firm size 2, but the rates have been very similar since 2013.
JOLTS relies on the federal Employer Identification Number (EIN) as reported with Unemployment Insurance (UI) filings to define a firm and to determine its size. Use of the EIN is consistent with other BLS data that are based on firm size and allows comparability with other size-based statistics. The definition for identifying firms by size differs from official JOLTS sample stratification methodology, which defines businesses by UI number and Reporting Unit Number (RUN) (defines a specific location). The EIN generally allows BLS to identify firms across state lines, while UI numbers are unique to each state where a firm does business.
To ensure adequate sample coverage, JOLTS groups firms into three employment size classes. These three JOLTS firm size classes are adjacent combinations of the 12 size classes defined by the Office of Management and Budget (OMB) as the standard for use by federal agencies. As a result, these size classes are comparable with many of the size class data sources that are available.
JOLTS sample selection methodology measures size using a 12-month maximum employment level. The size class of a firm is determined using the same 12-month maximum employment criterion. The firm size is assigned to all establishments in the firm each year with the new annual sample. Note that the JOLTS sample is selected at the UI/RUN (establishment) level, and therefore not all establishments in a firm are selected as part of the JOLTS sample. This methodology results in a stable sample distribution for three size classes. Previously-selected sample units are resized at the same time with the annual sample selection. This annual sizing approach creates “base-sized” estimates. For more information on sizing methodology, see the Business Employment Dynamics questions and answers page.
|Firm Size Class||EIN Max 12-Month Employment|
Since the JOLTS sample is stratified and selected on the basis of establishment size, it is necessary to post-stratify the sample to produce firm size estimates. The JOLTS frame is summarized using EIN by firm size to produce frame employment totals at the firm size level. As described earlier, each JOLTS establishment in the sample is assigned a firm size based on the 12-month maximum employment of the EIN that the establishment is part of. Using that firm size assignment, the JOLTS sample is then summarized by weighted sampled employment at the firm size level. The post-stratification factor is the ratio of the frame employment to weighted sampled employment at the firm size level. Note that this post-stratification technique assumes that the establishments in a firm that were not sampled have the same job openings, hires, and separations rates as the establishments in the firm that were sampled.
Individual firms may have establishments in several different industry groups. For example, a large firm that makes and distributes a product may have parts of its employment represented in the large size class of several industries like manufacturing, transportation and warehousing, and retail trade. For estimation purposes, data are placed in the industry of the individual establishments. That is, a firm's size class is assigned to each of the firm’s establishments in the establishment’s industry.
Due to the small sample size of approximately 16,000 establishments, JOLTS is publishing experimental firm size estimates for the total private industry only; no industry detail will be provided at this time.
The basic foundations for JOLTS size class estimates’ methodology are rooted directly in the methodology for producing JOLTS industry estimates. However, there are some differences in the methodology, as outlined below.
The weighted sample-based employment estimates are ratio-adjusted, or benchmarked, to independent population controls derived from the Bureau’s Current Employment Statistics (CES) program, which are more reliable due to a much larger sample size. Firm-size data are not available from CES. However, a proxy for CES firm size employment is derived using data from the Quarterly Census of Employment and Wages (QCEW). The QCEW distribution of firm size employment is applied to the CES industry/region estimates to derive an estimate for CES industry/region/firm size employment. This CES industry/region/firm size estimate is then used to benchmark JOLTS employment to the CES estimate.
The estimation cell is industry/firm size class. JOLTS experimental firm size-based estimates are calculated using the same Horvitz-Thompson estimator as the official published industry-based estimates.
The sampling weight for each sampled establishment is adjusted using the post-stratification factor discussed earlier. The cell estimate is the sum of the weighted responses in the cell times the cell nonresponse adjustment factor, times the cell calibration (benchmark) factor. The resulting estimates by industry/firm size class are then summed over industry to produce the aggregate firm size class estimates for total private.
To ensure that the sum of the three size class pieces is equal to the total private total, size class estimates are calibrated on a monthly basis.
The JOLTS Birth-Death model is the methodology used to estimate JOLTS data for the youngest establishments which are not included on the JOLTS sample frame. These younger establishments cannot be estimated from the sample because the sample frame is approximately 18 months old when the sample selected from it is first collected. Research on younger establishments on the QCEW has indicated that those establishments are systematically more active than older establishments with respect to employment change. They have higher employment growth and higher establishment death rates than older establishments. The current JOLTS sample has few, if any, responding young establishments. Consequently, the JOLTS Birth-Death model is needed to incorporate that growth and death activity. Since establishments older than 18 months are represented in the JOLTS sample, their hires and separations, including establishments that die, are captured through data collection and therefore do not need to be modeled.
The JOLTS Birth-Death model generates modeled hires and separations estimates for each establishment on the QCEW based on both the employment change of establishments on the QCEW and an estimate of “offsetting churn” (the level of hires and separations not accounted for by employment change) based on current JOLTS data. Modeled hires and separations estimates are generated for every record on the QCEW. The JOLTS Birth-Death model is essentially the modeled hires and separations estimates generated for the subset of establishments that are 18 months or younger. All records within this younger cohort are classified with respect to firm size. The JOLTS Birth-Death model firm size estimates are therefore the summary of modeled hires and separations by firm size.
JOLTS hires minus separations should be comparable to the CES net employment change. However, definitional differences between the two surveys as well as sampling and non-sampling error historically caused JOLTS to diverge from CES over time. To limit the divergence and to improve the quality of the JOLTS hires and separations series, JOLTS implemented the Monthly Alignment Method. JOLTS data are aligned at the industry/firm size level and then summed over all industries to create the total private level estimate. This procedure is independently performed on each characteristic—job openings; hires; total separations; quits; layoffs and discharges; and other separations.
Seasonal adjustment is the process of estimating and removing fluctuations caused by regularly occurring 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 characteristics (job openings; hires; total separations; quits; layoffs and discharges; and other separations) and employment for each firm size class are independently seasonally adjusted using the X-13 ARIMA-SEATS seasonal adjustment software. More information can be found on the Census website.
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 and discharges, and other separations) are computed in the same manner as the hires rate.
As the JOLTS firm-based size class estimates are newly-developed experimental data, variance estimates have not yet been developed.
JOLTS has produced size class estimates based on size of establishment since 2010. Those estimates are updated approximately every three months and are available upon request. Because the vast majority of firms are made up of single establishments, the behavior of the establishment-based estimates is close to the behavior of firm-based estimates. However, economic decision-making in multi-establishment firms is centralized at the firm level, making firm-based estimates more useful for studying behavior of firms of different sizes over time. Additionally, establishments that belong to a larger firm are thought to act differently than other establishments of the same size that do not belong to a larger firm. For the purpose of the JOLTS program, a firm is defined as all of the individual establishments that are assigned to a single Employer Identification Number (EIN). This number allows establishments to be grouped together under their parent firm. The size class data based on size of firm is being published for the first time in 2017 and is a different set of estimates than the establishment-based size class estimates.
Last Modified Date: September 13, 2018