• Overview
• Concepts
• Data sources
• Design
• Calculation
• Presentation
• History
  • Program identification and data publication
  • Major projects
  • Business Establishment List (BEL) Improvement Project
  • Electronic Data Interchange (EDI) Collection Center
  • Microdata/Macrodata (MIC/MAC) Project
  • SIC-to-NAICS conversion
  • Early BEA finance data
  • Business Employment Dynamics
  • QCEW Imputation Improvement Project (QIIP)
• More Info

• Methods Home
• About
• Programs
• Glossary
search

Handbook of Methods Quarterly Census of Employment and Wages History

Quarterly Census of Employment and Wages: History

The Quarterly Census of Employment and Wages (QCEW) is an administrative dataset guided by unemployment insurance (UI) coverage laws.^⁠1 These laws vary across states. They have evolved over time and continue to do so from state to state and year to year. We have more information about UI coverage historically. The following timeline presents important events in the history of the QCEW program.

Program identification and data publication

The QCEW program was initially named Employment Security Report 202, after the form on which data were collected. In 1998, to give some context to what the program produced, the name was changed to Covered Employment and Wages (CEW). The program became known as the Quarterly Census of Employment and Wages in 2002, to describe the mission more accurately.

The Standard Industrial Classification system (SIC) was used by the QCEW program to code establishments by industry until the fall of 2002, when the North American Industry Classification System (NAICS) was utilized in the publication of 2001 data.

In the late 1970s and early 1980s, special extracts and aggregations were provided to the public via mainframe computer tape and, later, on floppy disks. Alternatively, special print tables were produced with the information, as requested by customers. Data were subject to basic disclosure limitation procedures.

Following the early 1980s and into the late 1990s, the QCEW program followed a similar data pattern, but with more sophisticated disclosure limitation procedures. Some data requests were filled via CD, while more complicated requests required that BLS be reimbursed. During this timeframe, QCEW established subscription agreements with some of the larger private data companies. These companies would get updated files when they were available. Although data were scheduled to be released quarterly, they were actually released irregularly. On occasion, there were delays that pushed back file creation until more than one added quarter of data could be released at once.

From the late 1990s forward, requests that required reimbursement slowed and then ceased. More sophisticated disclosure limitation procedures were used. Early online tools offered annual average data for 1997 forward. In the early 2000s, a small number of downloadable files were posted on the BLS File Transfer Protocol site.

In 2014, QCEW data were made available as an open data product via comma-separated values (CSV) files on the program's web page. Open data, as described in OMB memorandum M-13-13 (PDF), refers to publicly available data structured in a way that enables the data to be fully discoverable and usable by end users. The CSV files are designed to allow third-party programmers, developers, and organizations to retrieve published QCEW data in CSV format. QCEW CSV files are organized by industry, by area, and by establishment size class.

Beginning with the publication of fourth quarter 2017 QCEW data, published in mid-2018, BLS accelerated the news release 2 weeks ahead of the normal full QCEW data update. BLS has changed this release schedule for two reasons. The first is to share QCEW data for all states faster. The second is to release the data before the Bureau of Economic Analysis (BEA) publishes quarterly gross domestic product (GDP) data, which uses the most recent QCEW wage data. BEA personal income estimates are heavily based on QCEW wages. For their earliest estimates, BEA uses Current Employment Statistics (CES) data to project wages. After BLS shares QCEW data with BEA, the GDP estimates are updated, replacing the CES-based forecast with QCEW wage data.

After reviewing the QCEW program’s procedures, BLS decided with the fourth quarter 2017 QCEW data to begin issuing the County Employment and Wages news release earlier.

The first QCEW release is a traditional news release, with a partial data update. This partial update will consist of the most recent monthly and quarterly levels presented in the County Employment and Wages news release. (See tables 1, 2, and 3.) The second QCEW release, consisting of the full QCEW dataset (current quarter, as well as any current year quarter, QCEW data), is published two weeks after the first QCEW (news) release.

Major projects

The QCEW program has undertaken several major projects in its brief history.

Business Establishment List (BEL) Improvement Project

BLS initiated the BEL Improvement Project to obtain information on multiple-establishment employers on a quarterly basis. The project involved changing the collection of data on employment and wages for multiple-establishment employers from a reporting-unit (county or industry total) basis to an individual-worksite (establishment) basis. As part of this change, the size criterion used to define multiple-establishment employers was lowered to include employers of smaller establishments. As a result, both the number of establishments and the number of multiple-establishment employers increased. The State Statistical Supplement forms that were previously used to collect county and industry summary-level data on multiple-establishment employers were replaced by the Multiple Worksite Report (MWR), a standardized form for use in all states.

The BEL project began in late 1988, using the Annual Refiling Survey (ARS). The survey obtained worksite identification information for existing multiple-establishment employers and identified those multiple-establishment employers which were previously categorized as single-unit employers. To capture ARS physical location address information for single-unit employers, the SIC Refiling Control System (SRCS) was modified in mid-1988 and the new version was provided to the states. The new, standardized MWR received OMB clearance and was mandated for state use beginning with data for the first quarter of 1991. (Note that, although state use of the form was mandatory, depending on state UI laws, it was not mandatory that all employers in the state use the form.) The MWR was intended to collect quarterly information on employment and wages from multiple-establishment employers.

The MWR was a major component of the BEL improvement project launched in 1989. The BEL moved the BLS sampling frame from a reporting-unit to an establishment basis. The Longitudinal Data Base (LDB) was established and later became part of the popular Business Employment Dynamics (BED) program. As mentioned earlier, the MWR replaced State Statistical Supplement forms. Several states had already collected establishment-level data, and their forms served as a model for the MWR. The initial quarter during which MWR establishment-level data were collected was the first quarter of 1991. The MWR enabled the QCEW program to be the first in BLS to develop a standardized magnetic-media reporting format. The MWR was implemented in September 1992.

Electronic Data Interchange (EDI) Collection Center

In 1995, BLS established the EDI Collection Center in Chicago, IL, to improve and to expedite the MWR collection process. Employers who complete the MWR for multiple-location businesses submit information on employment and wages via an electronic medium directly to the collection center, instead of submitting separate forms or files to each individual state agency. The collection center then edits the employers' data and forwards the data to the appropriate state agency. The EDI Collection Center is geared toward firms that meet certain size thresholds.

Electronic and web-based filing has reduced costs considerably for large and midsized employers. MWR web data collection started with data for the first quarter of 2007. The MWR electronic-reporting format is now routinely included in the products of prominent payroll- and tax-reporting software developers. Since 2012, BLS has successfully worked with a contractor to print, mail, receive, and scan returned MWR forms to state agencies, substantially reducing their workloads..

It is useful to examine some recent statistics regarding overall MWR data collection. In that regard, the numbers in the following tabulation, for the first quarter of 2019, exhibit a stable pattern that appears to be the usual pattern for MWR statistics:

As the data above show, 1.4 percent of employers completed the MWR for about 17 percent of national worksites and more than 41 percent of national employment. Moreover, 60 percent of all MWR employment came electronically, from either the EDI Center or MWRweb. Of this 60 percent, the EDI Center collected 24 percent of total MWR employment while MWRweb accounted for 36 percent.

Microdata/Macrodata (MIC/MAC) Project

Prior to 1989, states submitted only an enhanced file containing macrolevel employment and wage data. That meant that the highest level of detail received by the QCEW program was aggregate data cells made up of a combination of industry, county, and ownership codes.

In the first quarter of 1989, states began delivering a quarterly microdata file that included individual business establishment data. At the time, states were submitting nine separate deliverables per year: four quarterly microdata files, four quarterly macrodata files, and the annual Control File, providing Code Change Supplement data.

Development of the MIC/MAC project began in 1994. Teams consisting of BLS staff from both the national office and regional offices, as well as state staff members, were established to define the new processing requirements and deliverable standards. State and BLS systems were updated to reflect these major operational changes. The mission was to create a single quarterly file that would make possible the creation of macrolevel totals from the microdata. The project would promote more efficiency by eliminating duplication of work in reviewing and correcting the micro- and macrodata files. The project would also create uniformity in the micro- and macrolevel totals, as well as in the editing standards for the states and BLS. The first MIC/MAC files were submitted to BLS in the third quarter of 1997, providing for a more efficient work process and producing higher quality data.

SIC-to-NAICS conversion

The SIC system was established in the 1930s in order to classify establishments by their primary type of activity. The system was updated multiple times in its history of being utilized by the QCEW (formerly, ES-202) program; significant changes were made in 1967, 1972, and 1987. In July 1992, the Office of Management and Budget (OMB) established the Economic Classification Policy Committee, chaired by the Bureau of Economic Analysis (BEA) of the U.S. Department of Commerce, with representatives from the U.S. Bureau of the Census (now the U.S. Census Bureau) of the U.S. Department of Commerce and from BLS. The OMB charged the committee with conducting a “fresh slate” examination of economic classifications for statistical purposes and determining the desirability of developing a new industry classification system for the United States that would be based on a single economic concept. In 1997, NAICS was introduced, the product of a cooperative effort on the part of the official statistical agencies of the United States, Canada, and Mexico.

In the fall of 2002, the QCEW program published its 2001 data using NAICS 2002 classified industry data for the first time. Over several years prior to that milestone, state agencies had worked to convert more than 8 million establishments to NAICS. Along with the initial 1997 version of NAICS and the 2002 update, versions were released in 2007, 2012, and 2017, to reflect changes to the economy.

One of the major changes to industry coding that was introduced by NAICS was the sector called information (NAICS 51), which includes establishments producing and distributing information and cultural products, establishments providing the means to transmit or distribute these products, and establishments providing data or communications services, including data processing. Another new sector introduced by NAICS was management of companies and enterprises (NAICS 55).

The BLS conversion to NAICS was notable for its detail. An example may be seen in the construction (NAICS 23) sector. Within the specialty trade contractors (NAICS 238) subsector in the BLS implementation of NAICS, each base group in specialty trade contractors—for example, masonry contractors (NAICS 23814) or roofing contractors (NAICS 23816)—is further divided, on the basis of whether the establishment's predominant activity is in residential construction or nonresidential construction.^⁠2

Early BEA finance data

BEA is responsible for producing quarterly estimates of gross domestic product (GDP) and personal income (PI). A key component of the GDP and PI numbers is data on wages, from the QCEW program. Prior to 2008, BEA had difficulty estimating wages in a number of industries. The most difficult industry was finance (NAICS 52), because wage levels for that industry are greatly affected by irregular payments (for example, bonuses and stock options). These irregular payments are usually made in either the fourth or first quarter. In 2008, at BEA's request, BLS agreed to deliver first-quarter macrodata earlier than is customary, in order to enable BEA to adjust its estimates by means of QCEW total wage levels.

In early July of 2008, the 10 states with the largest proportion of national aggregate wages in the finance industry participated in a pilot test of accelerated macrodata delivery to BEA. The early-delivery pilot proved to be highly successful with major national users, such as the U.S. Department of the Treasury and the Council of Economic Advisers, which noted a decrease in revisions. On the basis of this success, BEA asked the QCEW program to provide early macrodata as a deliverable each first quarter, beginning with the first quarter of 2009. The BEA wage estimates that are being improved with QCEW data are national, not state, metro, or local, estimates.

Business Employment Dynamics

The Business Employment Dynamics (BED) program uses QCEW establishment data to produce a quarterly series of gross job gains and gross job losses for the entire economy. The program published its first release in 2003. The data covered the period from 1992 through 2002, with subsequent quarterly updates published since then. (For more on BED, visit the BLS website Business Employment Dynamics and the BLS Handbook of Methods for BED.)

QCEW Imputation Improvement Project (QIIP)

The economic situation in 2020, which reflects the impact of COVID-19, may not have been optimally captured with the old QCEW imputation method. In 2020, BLS explored three imputation (specific techniques of handling for non-response from active establishments) improvements. These improvements were accepted for production. The imputations were first used for the second published version of data for the first quarter of 2020 and for all versions of QCEW data beginning with the first published data for the second quarter of 2020.

The first improvement is a version of the ratio method of imputation described below.

BLS has conducted extensive research on alternative imputation methods for both employment and wages. The findings of the research indicate that current trends reported by similar businesses should be applied to nonrespondents. BLS defines the procedure for doing so as the ratio method. According to this method, the ratio for a particular estimation cell is computed as the sum of a current month’s reported employment (from similar businesses) divided by the sum of the previous month’s reported employment. To impute the current month’s employment for a nonrespondent, the ratio is then multiplied by the nonrespondent’s previous month’s employment. A similar procedure is applied to impute total quarterly wages.

The second improvement immediately identifies missing employers that are likely to have ceased operations. This eliminates imputation for many establishments that are out of business. The improvement uses counts of claims for unemployment benefits, aggregated by employer, to mechanically identify businesses that are likely to have ceased operations.

The third improvement is the use of benefit claims counts by state QCEW staff as a supplement to existing manual data review. Tables presenting the impact of the ratio method and the new identification of out-of-business establishments are available from the 2020 2nd quarter QCEW Improved Imputation Methodology Impact worksheet.