United States (1)

Title of the survey

Current Employment Statistics survey (CES)

Organization responsible

Department of Labor, Bureau of Labor Statistics (BLS)

Periodicity of the survey

Monthly.

Objectives of the survey

To provide timely, detailed and reliable estimates of employment, hours and earnings by industry. It is the first major economic indicator released each month (with estimates for the previous month). The statistics are used by policy makers, financial analysts, academia and the media. Additionally, CES estimates are used as inputs into other government statistics (e.g. Gross National Product, industrial production). The average hourly earnings series are commonly used by businesses for contract escalation purposes and in wage negotiations.

Main labour topics covered by the survey

Employment, earnings and hours of work.

Reference period

In private non-farm establishments and in state and local government: the pay period which includes the 12th of the month. In Federal Government establishments: the last day of the calendar month for employment.

Coverage of the survey

Geographical

The whole country.

Industrial

All branches of economic activity, except: agriculture, hunting, forestry and fishing (but agricultural services are included), the armed forces, and private household services.

Establishments

All types and sizes of non-agricultural establishments.

Persons

All employees in private establishments and civilian employees in government employment. Excluded are proprietors, the self-employed, unpaid volunteer or family workers, farm workers and domestic workers; military personnel in government employment, as well as employees of the Central Intelligence Agency and the National Security Agency.

Occupations

Occupational data are not collected.

Concepts and definitions

Employment

Employees are all persons on establishment payrolls who received pay for any part of the pay period which includes the 12th of the month. Salaried officers of corporations are included. For Federal Government establishments, employees are those persons who occupied positions on the last day of the calendar month. Intermittent workers are counted if they performed any service during the month. Persons on establishment payrolls who are on paid sick leave (when pay is received directly from the firm), on paid holiday, on paid vacation, or who work during a part of the pay period even though they are unemployed or on strike during the rest of the period, are counted as employed. Excluded are persons who are on layoff, on leave without pay, on strike for the entire period, or who were hired but have not yet reported during the period. The following categories of workers are separately identified:

Earnings

Data are collected on the total payroll of all full-time and part-time production, construction and non-supervisory workers, as defined above, who received pay for any part of the reference period. The payroll is reported before deductions of any kind (e.g. for old-age and unemployment insurance, group insurance, withholding tax, bonds or union dues). Included are pay for overtime, holidays, vacation and sick leave paid directly by the firm; other paid leave; bonuses paid at each pay period; incentive pay and commissions paid monthly. Excluded from total payroll are bonuses not paid at each pay period; other pay not earned in the pay period reported (e.g. retroactive pay); tips; pay advances; annual pay for unused leave; and the value of free rent, fuel, meals, or other payment in kind. Employee benefits (such as health and other types of insurance and contributions to retirement paid by the employer) are also excluded.

Wage/salary rates

Not relevant.

Hours of work

Data are collected on total hours paid for (including overtime), during the reference period, for production, construction and non-supervisory workers, as as defined above. Included are overtime hours, hours paid for stand-by or reporting time, hours not worked, but for which pay was received directly from the firm, for holidays, vacation, sick leave and other paid leave. Data are also collected separately on total overtime hours worked by production workers. These are the hours for which overtime premiums were paid because the hours were in excess of the number of hours of either the straight-time workday or the workweek during the reference period. Weekend and holiday hours are included only if overtime premiums were paid. Hours for which only shift differential, hazard, incentive or other similar types of premiums were paid are excluded.

International recommendations

The definition of payroll used in this survey complies with the international recommendations on regular earnings, with the following exception: the value of payments in kind (such as food and drink, fuel, clothing, etc.) and housing and rent allowances are excluded. The concept of hours of work used in this survey corresponds to hours paid for.

Classifications

Industrial

All data on employment, hours and earnings for the nation (beginning with August 1990) and for states and areas (beginning with January 1990) are classified according to the 1987 Standard Industrial Classification Manual (SIC). This classification is not directly linked to the International Standard Industrial Classification of all economic activities (ISIC), Rev.2, 1968. Establishments are classified into industries on the basis of their principal product or activity determined from information on annual sales volume. Since January 1980, this information is collected on a supplement to the quarterly unemployment insurance tax reports filed by employers. For an establishment making more than one product or engaging in more than one activity, the entire employment of the establishment is included under the industry indicated by the principal product or activity.

Occupational

Not relevant.

Others

By sex (the number of women employees is separately available). The states use the data to produce state and local area estimates.

Sample size and design

Statistical unit

This is the establishment, i.e. an economic unit, such as a factory, mine or store which produces goods or services. It is generally at a single location and engaged predominantly in one type of economic activity. Where a single location encompasses two or more distinct activities, these are treated as separate establishments, provided that separate payroll records are available.

Survey universe / sample frame

The universe file is the Unemployment Insurance (UI) register maintained by state Employment Security Agencies. Virtually all establishments must apply for the register when they are born and thus, the universe is constantly updated. It contains nearly seven million establishments. This file is augmented with information from other sources for those industries exempt from UI filing (for example, data on interstate railroads are obtained from the Interstate Commerce Commission, and Federal Government employment is obtained from the Office of Personnel Management).

Sample design

The sampling design is known as optimum allocation proportional to employment. Universe establishments are stratified first by region (state) then industry size. The sample comprises 1,703 basic cells. All large establishments (with 250 or more employees) fall into the sample with certainty. The number of required sample units for each smaller size class depends on the relative importance of the size class to the overall industry. For example, in a manufacturing industry with a high proportion of total employment concentrated in a small number of establishments, a larger percent of total employment is included in the sample. Consequently, for such industries, the sample design provides for a complete census of the large establishments, with a relatively few chosen from among the smaller establishments. For an industry in which a large proportion of total employment is concentrated in small establishments (e.g. in the trade and services divisions), the sample design again calls for inclusion of all large establishments, but also for a more substantial number of smaller ones. The CES sample covers over 390,000 reporting units with a total employment of 47 million. This represents a 4.4 per cent of the universe establishments and over one third of total universe employment. The sample is constantly augmented to maintain the sampling ratio, taking account of loss of reporting units due to business deaths or other attrition.

Field work

Data collection

Each month, the state agencies collect data on employment, payrolls and paid hours. Data are collected by mail and through automated computer-assisted method, or touch-tone self-response, voice recognition technology and electronic data interchange. A shuttle schedule (BLS form 790 entries) is used for mail respondents. It is submitted each month by the respondents, edited by the state agency and returned to the respondent for use again the following month. The state agencies forward the data, either on the schedules themselves or in machine-readable form, to BLS, Washington.

Survey questionnaire

The shuttle schedule consists of a one-page form accompanied with detailed instructions for entering data. The report forms are basically alike for each industry, but there are several variants tailored to the characteristics of different industries. They take the form of a table which includes questions on: for each month of the year. Thus, the design automatically shows the trends of the reported data and the relationship of the current data to the data for the previous months. The identification code and address of the reporting establishment is entered only once a year.

Substitution of sampling units

Sampling ratios are determined for each state-industry-size class. States are directed to solicit sample units based on these ratios for the given industry-size strata. After a period of non-response, a sampling unit will be dropped and replaced, if necessary to meet the sample requirements.

Data processing and editing

All schedules are edited by the state agencies each month to make sure that the data are correctly reported, and that they are consistent with the data reported by the establishment in earlier months and with the data reported by other establishments in the industry. States keypunch their sample data and telecommmunicate it to the BLS National Office. The data pass through a computerized edit and screening system: the first part tests for inconsistencies or impossibilities (e.g. incorrect industry code), and the second checks reporters' over-the-month changes relative to the rest of their industry sample, to detect outliers. All data which are initially rejected are reviewed by industry analysts who determine whether to accept the data or return it to the state for verification. Non-response is followed up by postcard or phone; verification is done by phone. States are directed to attach explanation codes to the sample data which undergo an unusual change. Some of these codes cause the sample data to be treated atypically for estimation purposes. For instance, atypical sample data is only allowed to represent itself, not other universe data which are outside the sample. If a report is not received from the establishment, but information is available from outside sources that the establishment was closed or some workers were off the payroll for the entire reference period, the estimates are adjusted accordingly.

Types of estimates

For employment purposes, part-time workers are counted the in the same way as full-time workers; they are most directly reflected in average weekly hours estimates, where they bring the industry average down. Workers not receiving pay during the entire pay period are not counted in the employment total, and so they do not affect hours and earnings data either. Workers absent for part of the pay period are counted in employment: their lower hours depress average weekly hours and their earnings carry less weight in computing average hourly earnings. At the basic estimating cell (industry, region and size), monthly data are obtained as follows: The unmodified sample average of average hourly earnings and average weekly hours are modified at the basic estimating cell level by a wedging technique designed to compensate for month-to-month changes in the sample of reporting establishments. The formula is as follows: Xc = (0.9Xp + 0.1xp) + (xc - xp) Where, Xc is the estimate for the current month, Xp is the estimate of average hourly earnings for the previous month, xp is the unmodified sample average for the previous month, xc is the unmodified sample average for the current month. The procedure reflected in this formula (i) uses matched sample data; (ii) tapers the estimate toward the sample average for the previous month of the current matched sample (xp) before applying the current month's change; and (iii) promotes continuity by heavily favoring the estimate for the previous month (Xp) when applying the numerical factors. At the aggregate industry level (division and, where stratified, industry), the monthly computation method is as follows: Average hourly earnings, excluding overtime, are computed by dividing the total production worker payroll for the industry group by the sum of total production worker hours and one-half of total overtime hours. Average hourly earnings, including lump-sum payments, are compiled for aircraft, and guided missiles and space vehicles manufacturing. The payroll data used to calculate this series include lump-sum payments made to production workers in lieu of general wage rate increases (such payments are excluded from the definition of gross payrolls used to calculate the other average hourly earnings series). For each sample establishment in these industries, the reported payroll data are adjusted to include a prorated portion of the lump-sum payment. Annual average estimates are computed as follows: Real earnings are expressed in constant dollars and are calculated from the earnings averages for the current month using a deflator derived from the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W). The reference year is 1982.

Construction of indices

Indexes of aggregate weekly hours and payrolls are prepared by dividing the current month's aggregates by the annual average aggregate for 1982. Indexes of diffusion of employment changes: these measure the dispersion among industries of the change in employment over the specified time span. Beginning with August 1990 data, the overall indexes are calculated from 356 seasonally adjusted employment series (three-digit industries) covering all non-farm payroll employment in the private sector. The manufacturing diffusion indexes are based on 139 three-digit industries.

Weighting of sample results

Estimates of employment are generated through an annual benchmark and monthly sample link procedure. Three basic steps are followed: Bias adjustment factors are then computed at the three-digit SIC level, and applied each month at the basic cell level, as part of the standard estimation procedures. Bias adjustment compensates for the under-representation of newly formed enterprises and other sources of non-sampling errors in the survey. Independent benchmarks are not available for the hours and earnings series. Consequently, the levels derive directly from the CES sample averages. However, a wedging technique is applied, which is designed to compensate for month-to-month changes in the sample of reporting establishments.

Adjustments

Non-response

The link relative technique avoids the need to make adjustments for missing reports. The reports with data for both the current and previous months form a matched sample and their over-the-month change is used to move the previous month's estimates.

Other bias

Bias adjustment factors are computed and applied as part of the standard estimation procedures. The main purpose of bias adjustment is to reduce the inability to capture, on a timely basis, employment generated by new firm births. Bias adjustment factors are computed quarterly, based on benchmark comparisons and modified by changes in rate of growth for the last quarter.

Use of benchmark data

Annual universe counts or benchmark levels are generated as of March of each year, in order to realign the sample-based employment totals for March of each year with total population counts for that date. Benchmark levels are generated primarily from administrative records on employees covered by Unemployment Insurance (UI) tax laws. Approximately 99 per cent of in-scope private employment is covered by UI. A benchmark for the remaining one per cent is constructed from alternate sources, primarily records from the Interstate Commerce Commission and the Social Security Administration. The March universe count replaces the initial estimates for that month. Monthly estimates for the year preceding the March benchmark are readjusted using a linear wedge back procedure, whereby the difference between the final benchmark level and the previously published March sample estimate is calculated and spread back across the previous 11 months. Estimates for the 11 months following the March benchmark are also recalculated each year, using the link relative procedure. Bias factors (see above) are updated to take into account the most recent experience of the estimates generated by the monthly sample versus the full universe counts derived from the UI. Following the revision of basic employment estimates, all other derivative series (e.g. production workers, average hourly earnings) are also recalculated. New seasonal adjustment factors are calculated and all data series for the previous five years are reseasonally adjusted, prior to full publication of all revised data in June of each year.

Seasonal variations

All series are seasonally adjusted using the multiplicative models under the X-11 ARIMA (Auto-Regressive Integrated Moving Average) method. Seasonal adjustment factors are computed and applied at component levels. For employment series, these are generally the 2-digit SIC levels. Seasonally adjusted totals are arithmetic aggregations for employment series and weighted averages of the seasonally adjusted data for hours and earnings series. Seasonally adjusted average weekly earnings are the product of seasonally adjusted average hourly earnings and seasonally adjusted average weekly hours. Average weekly earnings in constant dollars, seasonally adjusted, are obtained by dividing average weekly earnings, seasonally adjusted by the seasonally adjusted CPI-W, and multiplying by 100. Indexes of aggregate weekly hours, seasonally adjusted, are obtained by multiplying average weekly hours, seasonally adjusted, by production or non-supervisory workers, seasonally adjusted, and dividing by the 1982 annual average base.

Indicators of reliability of the estimates

Coverage of the sampling frame

The Unemployment Insurance registry covers about 99 per cent of the target population.

Sampling error / sampling variance

Several measures of the reliability of estimates are computed: Estimated standard errors for employment, hours and earnings: they are computed using the method of random groups and expressed as relative standard errors (standard error divided by the estimate). On the basis of sample data from March 1992 to March 1993, the relative standard errors of estimates for all non-farm industry divisions were as follows: Revisions between preliminary and final data: the root-mean-square error (i.e. the square root of the mean squared error) is computed, which measures the differences between final revised sample-based estimates (when nearly all the reports in the sample have been received) and first preliminary estimates based on less than the total sample. Other measures of error include benchmark revisions at the major industry division level and below, and computation of the standard error for differences between industries and times.

Non-response rate

Three estimates are made for a given month: two preliminary and one final, to allow for the inclusion of additional sample units. On average, 90 per cent of the sample units are received and matched by the third estimate, which represents about 91 per cent of sample employment.

Non-sampling errors

Benchmark revision is used as a measure of total sampling and non-sampling survey errors: over the past decade, percentage benchmark error has averaged 0.2 per cent, with a range from zero to 0.6 per cent.

Conformity with other sources

Each month, CES statistics are compared with the results of the Current Population Survey (CPS), and an analysis is made of the differences which may occur from each of these sources, account being taken of variations in definitions and coverage, source of information, methods of collection, and estimating procedures.

Available series

Published tables include: at the national level: at the state and area level: Seasonally adjusted data and historical data on employees by major industry, since 1943 and average hours and earnings of production or non-supervisory workers by major industry, since 1964, are also published monthly.

History of the survey

The first monthly studies of employment and payrolls by BLS began in 1915 and covered four manufacturing industries. Over the years, the survey was expanded to cover other manufacturing industries as well as non-manufacturing industries. Interest in employment statistics for States and areas also grew and cooperative arrangements between BLS and State agencies are now in effect with all 50 states, the District of Columbia, Puerto Rico and the Virgin Islands. The major changes that have occurred have involved industry classification. The latest conversion occurred in September 1990 when BLS switched from the 1972 SIC to the 1987 SIC system. Prior to the 1972 SIC, the survey was based on the 1967 SIC. In all cases, attempts have been made to insure the continuity of the time series and the history of those affected by SIC revisions has been reconstructed when possible. The published industry detail for the service-producing sector has been expanded, bringing it into parity with the goods-producing sector. The CES programme also has an increasing number of large sample units which report their data automatically via phone (Touch-tone data entry) rather than through the mail, in order to facilitate non-response prompting, improve the response rate for the first estimates, and thus minimize revisions.

Documentation

US Department of Labor, Bureau of Labor Statistics: Employment and Earnings (monthly, Washington). Preliminary estimates and summary data are available about five weeks after the reference week of the survey. Preliminary estimates for the full industry detail, based on about 80 per cent of the sample, are in the following month's issue. Final (pre-benchmarked) figures are issued one month later, i.e. three months after the survey reference period. Benchmark revisions are usually published each year in the June issue of Employment and Earnings. idem: Monthly Labor Review (monthly, ibid.). Historical national statistics (monthly data and annual averages) appear in Employment, Hours and Earnings, United States, 1909-90 (Bulletin 2370); updated by the Supplement to Employment and Earnings (annual, ibid.). Methodological information is published in each monthly issue of Employment and Earnings, and is also available in: idem: BLS Handbook of Methods, Bulletin No. 2414 (September 1992). Employment, hours and earnings data are available in machine-readable form and on diskettes on a cost-for-service basis.

Confidentiality / Reliability criteria

Confidentiality considerations prevent the disclosure of establishments which are in the CES survey, as well as establishment-specific data or any estimates dominated by a single reporter. Some of the industry estimates are unpublished because they do not meet BLS standards for accuracy and reliability.

Other information

Data supplied to the ILO for publication

The following data derived from the CES survey are published in the Yearbook of Labour Statistics: The corresponding monthly series of paid employment (all employees), average hours paid for and average hourly earnings of production and related workers, in non-agricultural activities and manufacturing, are published in the Bulletin of Labour Statistics.