To receive a copy of Regional and State Employment and Unemployment (LAUS) or Metropolitan Area Employment and Unemployment (METRO) monthly press releases via e-mail, link to the BLS News Release Subscription page at www.bls.gov/bls/list.htm and subscribe to the e-mail service..
Over the course of a year, the size of a state's employment level undergoes sharp fluctuations due to such seasonal events as changes in weather, reduced or expanded production, harvests, major holidays, and the opening and closing of schools. Because these seasonal events follow a more or less regular pattern each year, their influence on statistical trends can be eliminated by adjusting the statistics from month-to-month. These adjustments make it easier to observe the cyclical and other nonseasonal movements in the series. In evaluating changes in seasonally adjusted series, it is important to note that seasonal adjustment is merely an approximation based on past experience. Seasonally adjusted estimates have a broader margin of possible error than the original data on which they are based, because they are subject not only to sampling and other errors but are also affected by the uncertainties of the seasonal adjustment process itself. Employment data are seasonally adjusted with a procedure called X-12-ARIMA. State major industry division employment are available electronically via the Internet and are published monthly in Employment and Earnings.
For a further examination of our seasonal adjustment process please consult the following resources:
Berger, Frank and Keith Phillips, The Disappearing January Blip and Other State Employment Mysteries, Working Paper 94-03, Federal Reserve Bank of Dallas, February 1994.
Dagum, Estela Bee. The X-11 ARIMA Seasonal Adjustment Method. Ottawa, Statistics Canada, January 1983, Statistics Canada Catalogue No. 12-564E.
Shipp, Kenneth and Thomas J. Sullivan, “Using X-11 ARIMA to Seasonally Adjust State Level Industry Employment Data in the Current Employment Statistics Program,” Proceedings of the ASA Business and Economic Statistics Section, 1992.
Stuart Scott, George Stamas, Thomas J. Sullivan, and Paul Chester, Seasonal Adjustment of Hybrid Economic Time Series, American Statistical Association, Toronto, Canada, 1994.
Employment is the total number of persons on establishment payrolls employed full or part time who received pay for any part of the pay period which includes the 12th day of the month. Temporary and intermittent employees are included, as are any workers who are on paid sick leave, on paid holiday, or who work during only part of the specified pay period. A striking worker who only works a small portion of the survey period, and is paid, would be included as employed under the CES definitions. Persons on the payroll of more than one establishment are counted in each establishment. Data exclude proprietors, self-employed, unpaid family or volunteer workers, farm workers, and domestic workers. Persons on layoff the entire pay period, on leave without pay, on strike for the entire period or who have not yet reported for work are not counted as employed. Government employment covers only civilian workers.
A sample establishment in the CES survey is an economic unit, such as a factory, which produces goods or services. It is generally at a single location and engaged predominantly in one type of economic activity. Establishments reporting on the schedule (form BLS 790) are classified into industries based on their principal product or activity determined from information on annual sales volume. This industry classification, based on the 2012 North American Industry Classification System (NAICS) Manual, is collected on a supplement to the quarterly unemployment insurance tax reports filed by each employer. For an establishment making more than one product, the entire employment is included under the industry of the principal product or activity.
Geographic hours and earnings CES estimates are available only for production workers primarily in manufacturing industries. Because not all sample respondents report production worker, hours, and earnings data, insufficient sample exists to make corresponding industry estimates of average weekly hours and average hourly earnings outside of manufacturing at the statewide level. National estimates of average weekly hours and average hourly earnings are made for the private sector, with detail for about 850 private industries as well as for overtime hours in manufacturing.
Hours and earnings are derived from reports of gross payrolls and corresponding paid hours for production workers, construction workers, or nonsupervisory workers in the service sector. The payroll for workers covered by the CES survey is reported before deductions of any kind, e.g., for old-age and unemployment insurance, withholding tax, union dues or retirement plans. Included in the payroll reports is pay for overtime, vacations, holidays and sick leave paid directly by the firm. Bonuses, commissions, and other non-wage cash payments are excluded unless they are earned and paid regularly—at least once a month. Employee benefits paid by the employer, tips, and payments in kind also are excluded.
Total hours during the pay period include all hours worked (including overtime hours) and hours paid for holidays, vacations, and sick leave. Total hours differs from the concept of scheduled hours worked. The average weekly hours reflects effects of numerous factors such as unpaid absenteeism, labor turnover, part-time work, strikes, and fluctuations in work schedules for economic reasons. Overtime hours in manufacturing are collected where overtime premiums were paid if hours were in excess of the number of straight time hours in a workday or workweek.
Each month BLS collects data on employment, hours, and earnings from a sample of about 554,000 nonfarm establishments which employ nearly 40 percent of the total nonfarm population. All establishments with 1000 employees or more are asked to participate in the survey along with a representative sample of smaller establishments. Sample respondents extract the requested data from their payroll records, which must be maintained for a variety of tax and accounting purposes. Initially, data were collected primarily by mail until recent BLS initiatives in collection methodology increased the use of electronic media. Now, web collection, phone collection, touch-tone self response, computer-assisted interviews, fax technology, and voice recognition are also being used to obtain higher and faster response rates.
Data submitted on the schedules are used by BLS analysts in developing statewide and major metropolitan area estimates. All states' samples are combined to form a collective sample for developing national industry estimates. Statewide samples range from nearly 58,000 sample units in California to about 1,400 units in smaller states. It should be noted that BLS estimation procedures are designed to produce accurate data for each individual state. BLS independently develops the national employment series and does not force state estimates to sum to national totals nor vice versa. Because each state series is subject to larger sampling and non-sampling errors than the national series, summing them cumulates individual state levels errors and can cause significant distortions at an aggregate level. Due to these statistical limitations, BLS does not compile a “sum of states” employment series and cautions users that such a series is subject to a relatively large and volatile error structure.
Employment estimates are made at the publication cell level and aggregated upward to broader levels of industry detail. A minimum guaranteed publication structure has been defined for all States and MSAs. The structure consists of “expanded” supersectors, which break Manufacturing; Trade, Transportation, and Utilities; and Government into further publication detail. The guaranteed publication cells aggregate to the summary cells of goods-producing, service-providing, total private, and total nonfarm employment. All other published series had to pass a minimum sufficiency test of at least 30 unique unemployment insurance (UI) accounts in its sample, or a minimum universe employment count of 3,000 with at least 50 percent covered by the sample. The series were tested using employment data from the Covered Employment and Wages program (CEW, or ES-202). See www.bls.gov/sae/saenaics.htm#guaranteed for more information.
Guaranteed industries that do not pass the minimum sufficiency test are estimated using a regression model. The CES Small Domain Model (SDM) is a Weighted Least Squares model with three employment inputs: (1) an estimate based on available CES sample for that series, (2) an ARIMA projection based on trend from 10 years of historical data, and (3) an estimate “borrowed” from the Statewide series for that industry. In addition to the guaranteed industries, Sectors may be modeled at the Statewide level. Approximately 44 percent of State and area CES series are model-based.
For each non-summary cell a total level of benchmark employment is obtained for a specific month (usually March). The sample data from reporters who responded for consecutive months provides a link relative sample ratio. This ratio is applied to the benchmark employment month to produce an April employment estimate. This process continues each month until the next annual benchmark cycle when estimates are replaced with population data. States also use a net birth/death factor to supplement the link relative estimator in the monthly estimation process. Birth/death factors are used to compensate for the inability to capture the entry of new firms into the sample, as well as the exit of firms that went out of business from the sample, on a timely basis.
For example, assume the benchmark level was 50,000 in March. The sample, composed of 50 establishments which reported both months had 25,000 in March and 26,000 in April, a 4 percent increase. Also, there is an April birth/death factor of 300. To derive the April estimate, the change of these identical establishments reported is applied to the March benchmark level in the form of a sample ratio, then the birth/death factor is applied to this number: (50,000 x 26,000/25,000) + 300 = 52,300.
To control potential survey error, the estimates are benchmarked annually to universe counts derived from administrative files of employees covered by unemployment insurance (UI). Original sample-based estimates are replaced with benchmark data from the previous year through at least March of the benchmark year. In the current 2013 benchmark, the estimates from April 2012 to September 2013 were replaced with UI-based universe counts. For more info, see our Benchmark article at www.bls.gov/sae/benchmark2014.pdf. Once the new September 2013 level was determined, the subsequent estimates were recalculated by applying the appropriate sample links to the new levels. These links may differ slightly from those used to derive the original estimates, because they account for late reporters. The entire period from October 2013 forward is referred to as the post-benchmark projection period. This process was completed and the revised data were released with the January 2014 estimates.
Yes, estimates are revised in the following manner.
Initial monthly estimates are calculated from an incomplete sample and are subject to revision in the subsequent month when more sample data are available. Revisions at the total nonfarm levels for preliminary statewide employment are generally small. For more info on preliminary-to-final revisions, please visit www.bls.gov/sae/saerevtxt.htm.Final-to-Benchmark Estimates
“Final” estimates are subject to annual benchmarks of universe counts of employment derived from the unemployment insurance (UI) reports from employers. The average absolute benchmark revision at the state total nonfarm level was 0.4 percent in March 2013. The average absolute revision from 2008 to 2012 was 0.6 percent. The range of the percentage revision for the states at the total nonfarm level was from -0.7 to 2.9 percent in March 2013. A positive revision indicates that the benchmark value is greater than the sample-based estimate and vice-versa.
Thirty-one states and the District of Columbia revised total nonfarm payroll employment upward, while 19 states had downward revisions. Click here for more information on the benchmark.
As mentioned earlier, geographic data are subject to changes in administrative mandates for revising NAICS and metropolitan area definitions. The CES program has consistently attempted to maintain industry and area time series, particularly at guaranteed publication levels, where data and resources permit. Data are annotated where reconstruction of time series is not possible.
Advantages of CES Geographic Data
CES data are a coincident economic indicator and are often cited in national and local newspapers, magazines, and reports. This press generates enthusiasm, curiosity and a wealth of outside material for supplementary reading. The College of Business Administration at the University of South Carolina uses seasonally adjusted employment as an indicator of current employment trends in South Carolina. The regional Federal Reserve Banks use CES data in easy-to-understand economic applications. For example, the edition of the Southwest Economy from the Federal Reserve Bank of Dallas used employment and unemployment data in two different articles: one explaining the Phillips curve and another describing the changing job market. Students and faculty can write the regional FRBs to be placed on their mailing lists. The Philadelphia, Dallas, Boston, Cleveland, and San Francisco FRBs provide excellent articles for undergraduate students.
CES data are tangible and versatile. Employment, hours, and earnings data can be used to study abstract economic concepts which students can more easily comprehend with the use of data. Students often need help in seeing how formal models can be used to explain the real world economy. Business cycles, the effects of shocks in the economy, and the impact of policy changes are examples of concepts that are more readily understood when using CES data. Also, combined with data from other sources, such as output data from the national accounts, they can be used to compute productivity and other measures. Primarily, the concept of employment is easy to comprehend, which permits a wide range of study and understanding by graduate and undergraduate students, policy makers, and business people. Data can be used for projects in labor economics, time series analysis, business cycle theory, statistics, geography, urban planning, and public policies.
CES data invite comparisons and analysis. CES data provides complete coverage and consistently derived methodology at the state and area levels for employment in major industries allowing for interstate and inter-area comparisons using CES data alone or in conjunction with other economic data. They allow one to compare growth patterns across states and regions. One can relate cyclical changes to geographic employment changes. For example the 1990-91 recession did not affect states and regions equally or at the same time. Employment declines started in the Northeast and spread along the Atlantic and Pacific coasts. The Midwest was largely unaffected. These diverse movements among states show how the mixture of industry, migration, and public policies affect employment. For this type of study, CES data can be combined with and compared to census migration data, immigration data, and public policy data that affect economic activity.
CES data are affordable. They are collected, tabulated, and distributed as part of the BLS and States' mission to provide economic data to policy makers, business, labor, and the public. Subscriptions are inexpensive and data on Internet are free. Since CES data are time series data, forecasters are able to depend on a consistent series to use in their modeling applications without incurring excessive costs.
Users should be aware of the intricate revision process which the CES estimates undergo. Preliminary monthly, final monthly, post benchmark projection, and final benchmark data are constructed for each monthly estimate. Analysis using estimates before they are final benchmarked estimates is affected by subsequent revisions.
Users of time series CES data should also review the entire time-series file to note any NAICS or MSA administrative breaks where reconstruction of series was not possible. Breaks will only be noted on the month where the time series break occurs. For example, a comparison of total nonfarm employment for the Washington D.C. metropolitan area between 1980 and 2003 actually involves multiple definitions of the official metropolitan area.
As mentioned earlier, the CES national estimates are independently produced and are not an aggregation of statewide data. Therefore users cannot disaggregate or compare CES national economic movements to state, regional, or metropolitan area CES estimates.
CES data are not to be confused with data from the Current Population Survey (CPS) which is a household survey. The CES survey counts jobs; the CPS counts people. A worker with two jobs is counted twice in the CES but only once in the CPS.
Geographic hours and earnings data from the CES are limited in industry coverage and scope. The only extensive industry coverage is in manufacturing. CES hours and earnings data are also limited to money wages of production workers in manufacturing. Researchers looking at total labor costs and total compensation should be aware of these limitations.
If you have a question related to the Current Employment Statistics Survey dealing with State and area data, feel free to send an e-mail.
Barth, Molly E., “Revisions to the Current Employment Statistics State and Area Estimates Effective January 2003,” Employment and Earnings, March 2003.
Barth, Molly E., “Recent changes in the State and Metropolitan Area CES survey,” Monthly Labor Review, June 2003.
BLS Handbook of Methods, April 1997.
Current Employment Statistics State Operating Manual, October 1989 (with annual updates).
Dahlin, Brian, “Revisions in State Establishment-Based Employment Estimates Effective January 2003,” Employment and Earnings, May 2003.
Dagum, Estela Bee. The X-11 ARIMA Seasonal Adjustment Method. Ottawa, Statistics Canada, January 1983, Statistics Canada Catalogue No. 12-564E.
Employment and Earnings, monthly.
Employment, Hours, and Earnings, United States, 1909-90, volumes I and II, Bulletin 3270, March 1991, annual supplement, August 1992.
Employment, Hours, and Earnings, States and Areas , 1987-94, September 1994.
Green, Gloria P., “Comparing Employment Estimates from Household and Payroll Surveys,” Monthly Labor Review, December 1969.
Kropf, Jurgen, Strifas, Sharon and Traetow, Monica, Accounting for Business Births and Deaths in CES: Bias vs. Net Birth/Death Modeling, 2002.
Manual on Series Available and Estimating Methods, BLS Current Employment Statistics Program, March 1994.
Morisi, Teresa L. “Recent changes in the National CES survey,” Monthly Labor Review, June 2003.
National Commission on Employment and Unemployment Statistics. Counting the Labor Force, 1979.
Last Modified Date: March 18, 2014