Questions and answers about methods underlying the monthly employment reportPayroll survey questions Payroll survey questions
No, the CES program covers nonfarm wage and salary jobs. It excludes workers in agriculture and private households and the self-employed. CES is a survey of business establishments whose goal is to measure wage and salary employees on payrolls. The sampling frame for the survey (derived from UI tax records) does not cover much of the nation’s self-employment. To get a comprehensive measure of self-employment it is preferable to survey households. The Current Population Survey (CPS), also known as the household survey, generates estimates of the self-employed. The CPS and CES are both sample-based surveys and both have strengths and limitations. The CES employment series has a much smaller margin of error on the measurement of month-to-month change than the CPS because of its much larger sample size. An over-the-month employment change of 104,000 is statistically significant in the CES, while the threshold for a statistically significant change in the CPS is 436,000. However the CPS has a more expansive scope than the CES as it includes self-employed and unpaid family workers, agricultural workers and private household workers that are excluded from CES. No, the CES has a probability-based sample design that represents all industries appropriately. About 5 percent of the establishments in the CES sample are in manufacturing. Yes; about 40 percent of the CES sample is composed of business establishments with fewer than 20 employees. The CES sample is designed to maximize the reliability of the total nonfarm employment estimate; firms from all size classes and industries are appropriately sampled to achieve that goal. Yes, for current monthly estimates BLS makes a model-based adjustment to account for the net employment change generated by business births and deaths. The CES uses modeling rather than sampling because the survey is not able to immediately bring new businesses into the sample. There is an unavoidable lag between the birth of a new firm and its appearing on the sampling frame and being available for selection. BLS does sample new businesses twice a year, but with a lag. CES revises to improve its data series, by incorporating additional information that is not available at the time of initial estimate publication. CES revises its initial monthly estimates twice, in the immediately succeeding months, to incorporate additional sample receipts from respondents in the survey. On an annual basis CES incorporates a benchmark revision that re-anchors estimates to complete employment counts available from unemployment insurance tax records. The benchmark helps to control for sampling and modeling error in the estimates.
Household survey questions
Receipt of unemployment benefits is not part of the concept of unemployment as defined by the CPS. In order to be classified as unemployed, one must be without a job, able to take a job, and have actively searched for a job during the prior 4 weeks. No questions are asked regarding unemployment benefits. Since persons must have been engaged in an active job search to be classified as unemployed, those who have stopped looking for a job are not included. The CPS does provide estimates of persons outside the labor force who want a job, including discouraged workers, who are those who are not looking because they believe no jobs are available for them. The CPS also publishes each month alternative measures of labor utilization that take into account groups of workers, such as discouraged workers, who are not counted among the unemployed. Both the CPS and CES surveys have strengths and limitations. The CES employment series has a much larger sample than the CPS, and therefore a much smaller margin of error. The CES series are benchmarked yearly to universe counts. CPS estimates, in contrast, are not directly benchmarked to a universe count, but are controlled to estimates of the civilian noninstitutional population. The CPS does have a much broader scope than the CES, and includes workers that the CES does not cover, such as the self employed, unpaid family workers, agriculture workers, private household workers, and those on unpaid absences. Since decennial censuses occur just once very 10 years, it is difficult to know how accurately the population estimates reflect the true population in the years in between censuses. There are limitations to the population control estimates given the difficulties in estimating the number of illegal immigrants and other factors. The Census Bureau continues to look for ways to improve the accuracy of the intercensal population estimates and controls. The BLS cannot determine to what extent illegal immigrants are reflected in the household survey, though it is likely that it does include some illegal immigrants. The household survey does not include questions to identify one’s legal status, but it does include questions about whether respondents were born outside the U.S. The CPS shows that foreign-born workers (whether legal or illegal) accounted for about 16 percent of the labor force in 2007 and about 48 percent of the net increase in the labor force from 2000 to 2007.
Questions about differences between the payroll and household survey
Estimates of unincorporated self employed from the household survey have shown no substantial growth in recent years. Any number of self-employed persons work as independent contractors. Some independent contractors are not reported correctly as self employed in the household survey, but rather as wage and salary workers. According to BLS studies of estimates of independent contractors from special supplemental surveys to the CPS, about 12 to 15 percent of independent contractors are mistakenly reported as wage and salary workers. The household survey may be overstating the number of wage and salary workers through some misclassification of independent contractors. For more information about classification of independent contractors in the household survey, see “Understanding the employment measures from the CPS and CES survey,” from the February 2006 Monthly Labor Review, at www.bls.gov/opub/mlr/2006/02/art2full.pdf. The effect of any growth in undocumented workers on the gap between the household and payroll surveys is unknown. While it is likely that both surveys include at least some illegal immigrants, neither survey tries to identify the legal status of workers. The household survey does include questions about whether members of the household were born outside of the U.S. CPS data show that foreign-born workers accounted for about 16 percent of the labor force in 2007 and about 48 percent of the net increase in the labor force from 2000 to 2007. For the latest BLS data on foreign-born workers, see Labor Force Characteristics of Foreign-Born Workers Since decennial censuses occur just once very 10 years, there is no way to know how accurately the population estimates reflect the true population in the years in between censuses. The population controls were understated in the 1990s. When population controls for Census 2000 were introduced, they raised the level of CPS employment by about 1.6 million, explaining over one-third of the 4.5 million cumulative gap between CPS and CES in the late 1990s. There are limitations to the population control estimates given the difficulties in estimating the number of illegal immigrants and other factors. The Census Bureau continues to look for ways to improve the accuracy of the intercensal population estimates and controls. For further discussion on population controls and possible overstating/understating of CPS employment, see “Understanding the employment measures from the CPS and CES survey,” from the February 2006 Monthly Labor Review, at www.bls.gov/opub/mlr/2006/02/art2full.pdf, and Examining the discrepancy in employment growth between the CPS and CES, at www.bls.gov/bls/fesacp2101703.pdf (pp. 9-14)
Back to Monthly Employment Situation Report: Quick Guide to Methods and Measurement Issues
Last Modified Date: April 18, 2008 |
