Skip Main Navigation Skip To Navigation Content

Labor Force Statistics






Guidance



Guidance on Differences in Employment
and Unemployment Estimates from Different Sources

August 23, 2004




The federal government reports employment and unemployment (labor force) estimates from several major surveys and programs:

 

   Current Population Survey (CPS)

 

   American Community Survey (ACS)

 

   Local Area Unemployment Statistics (LAUS) Program

 

   Current Employment Statistics (CES) Program

 

   Survey of Income and Program Participation (SIPP)

 

These reports are compiled from three major sources:

 

   Household surveys, such as the CPS and the ACS, provide data (sometimes referred to as household data) that pertain to individuals, families, and households, and relate to where people live (as opposed to where they work).

 

   Reports based on information from employers, such as the CES, provide data from employers or establishments that pertain to jobs (people on a payroll) and relate to where those jobs are located.

 

   Records based on information from unemployment insurance (UI) records pertain to people who are eligible for UI benefits based on program requirements, including those who file claims, and relate to the place of insurance coverage of the establishment.


Estimates of employment and unemployment from a particular program may differ from those of other programs because of variations in definitions, coverage, methods of collection, reference periods, and estimation procedures. These variations, in turn, arise from differences in the programs’ purposes. The various estimates of employment and unemployment generally complement one another, with each providing a different type of information that the others cannot suitably supply. This multiplicity of sources makes it important to understand when it is appropriate to use the data from each source.


HIGHLIGHTS


Current Population Survey (CPS)


Because of its detailed questionnaire and its interviewing staff trained to explain labor force concepts and answer questions, the CPS is a high quality source of information used to produce the official monthly estimates of employment, unemployment, and the unemployment rate for the nation and states. It is also a source of information on other labor force topics such as actual hours of work and duration of unemployment.



American Community Survey (ACS)


Since 2000, while in its testing phase, the ACS has provided annual estimates of employment and unemployment (essentially for all areas of 250,000 population or more). The sample size of approximately 800,000 addresses per year make it useful for subnational analyses. When fully implemented (planned for 2005), the ACS will have an annual sample size of 3 million addresses nationwide. The ACS uses the same labor force questions as Census 2000.


Because of its large sample size, the ACS will have advantages over the CPS in producing estimates in the following circumstances:

 

   to characterize small geographic areas for which CPS (or Local Area Unemployment Statistics Program) estimates are not available, and for comparisons among such areas and between such areas and larger ones;

 

   to provide information on socioeconomic characteristics of the labor force that are not collected in the CPS, or for geographic areas below the level for which the CPS can provide this information;

 

   to produce tabulations of finely detailed categories, or extensive cross-tabulations of multiple characteristics of the labor force for any geographic area, including the nation, for which the CPS sample size is insufficient to produce reliable estimates;

 

   to study rare characteristics of common population groups, or characteristics of uncommon population groups;



Local Area Unemployment Statistics (LAUS) Program


The LAUS program is a cooperative federal-state program that produces monthly and annual estimates of employment, unemployment, and the unemployment rate for over 7,000 geographic areas. The CPS data are a key input to the LAUS program methodology and have a significant impact on the estimates from the program. Together, the CPS and LAUS program estimates provide a consistent historical time series for employment and unemployment data at the national and state level.



Survey of Income and Program Participation (SIPP)


The Survey of Income and Program Participation (SIPP) is a longitudinal survey. It is useful mainly for examining the dynamics of employment change (including spells of unemployment) for the same individuals over time.



Current Employment Statistics (CES) Program


The CES program is an excellent source of information on employment, hours worked, and hourly and weekly earnings as reported by a sample of almost 400,000 establishments. Employment, hours worked, and earnings data are based on payroll reports. This survey estimates the number and characteristics of jobs held, not the number of people employed.



DETAILED COMPARISON OF SOURCES AND THEIR ESTIMATES


Household Data


The Current Population Survey (CPS), the American Community Survey (ACS), and the Local Area Unemployment Statistics (LAUS) Program


The CPS is a high-quality sample survey of the population 16 years and over, conducted each month by the Census Bureau for the Bureau of Labor Statistics (BLS). It provides accurate and timely data needed for swift economic decision making by the Administration, Congress, and the Federal Reserve Board. The CPS has been the source of the official estimates of employment and unemployment for the nation for more than 50 years. The survey has been greatly expanded and improved over the years, but the basic concepts of employment and unemployment – reviewed periodically by high-ranking commissions – have remained substantially unaltered.


The CPS is specifically designed to be the official source of monthly estimates of employment and unemployment for the United States, and of annual-average estimates of these measurements for all states. It also publishes annual-average estimates for 50 large metropolitan areas (MAs) and 17 cities, but its sample is not large enough to provide reliable data for other areas. The CPS provides information on the detailed socioeconomic characteristics of the labor force for the nation, states, and published areas.


The ACS, when fully implemented, will be the largest household survey in the United States. It will provide direct annual estimates of employment and unemployment for the nation and states, and for counties and other governmental units of 65,000 or more population. For smaller governmental units, it will annually provide 3-year or 5-year moving averages. Eventually, the ACS will be able to measure changes over time for small areas and population groups. The ACS will provide a wealth of data on socioeconomic characteristics that can be related to its labor force estimates.


Monthly labor force and unemployment estimates for states, labor market areas (LMAs), and other areas covered under federal assistance programs are developed by state employment security agencies and the BLS under the federal-state cooperative LAUS program. LAUS program data are pegged to the place of residence and labor force concepts of the CPS. The LAUS program uses several methods to produce estimates for more than 7,000 geographic areas, including all counties and cities of 25,000 or more, and all cities and towns in New England. 1 Estimates are available on a monthly and annual basis, with preliminary estimates published within 5 weeks after the reference period. No socioeconomic data are associated with this series.


Definitional and procedural differences between the CPS and the ACS are important to consider when using or comparing estimates. These differences also affect the use and comparison of LAUS and ACS estimates, since the CPS is the key input to the LAUS models of the employment-to-population ratio and the unemployment rate for states and other large areas, as well as having a significant impact on the labor force estimates for the other 7,000 substate areas.


The remainder of this section describes differences among the household-based estimates.


The CPS and LAUS program provide monthly estimates. The ACS provides annual information about the labor force for small areas (essentially all cities, counties, and metropolitan areas of 250,000 population or more during the testing phase; governmental units as small as 65,000 population when fully implemented, and all areas as small as census tracts beginning in 2010), as well as for the nation and the states.


The CPS provides a monthly snapshot of national-level employment and unemployment statistics with less than a 3-week turnaround between the end of data collection in any one month and release of the statistics for that month by the BLS. The ACS does not produce monthly estimates because it has been optimized instead to produce accurate estimates for geographic areas as small as census tracts and block groups.


The CPS computes annual averages of employment and unemployment statistics at the state level and for the New York and Los Angeles metropolitan areas. The LAUS program develops monthly estimates for substate areas including all metropolitan areas, small labor market areas, counties, cities of 25,000 or more, and all cities and towns in New England. The ACS will have a large enough sample to provide annual estimates for much smaller geographic areas, with those for areas with populations under 65,000 being 3-year or 5-year moving averages.


The CPS asks a more detailed series of questions than the ACS does about labor force participation in order to obtain the accuracy required for the official labor force estimates. The CPS data are collected by field staff via personal interviews that allow for followup questions to clarify complicated concepts or to probe for information to implement the official definition. 2 The Census 2000 long form, which used the same question sequence as the ACS, produced generally higher estimates of unemployment than the CPS.


Because the CPS asks detailed questions and has an overlapping sample design, it provides an estimate most consistent with the official definition of the unemployment rate, and is able to detect month-to-month changes in this key economic indicator with considerable precision, based on a sample size of 73,000 per month for labor force statistics.


The ACS sample size is 800,000 addresses annually during the testing phase and will be 3 million addresses annually when the survey is fully implemented (planned for 2005). The large ACS sample size provides unrivaled coverage and precision of measurement for small areas, sufficient cases to study relatively rare events, and the ability to provide cross-classifications of multiple characteristics.


The reference period for the CPS labor force estimates for a given month is the calendar week including the 12th day of the month. Employment data about the week of the 12th is collected during the week of the 19th. The week of the 12th was selected as the reference week for employment to minimize the effect of holidays and other seasonal variation. Annual CPS estimates are obtained by averaging the twelve monthly estimates. The reference period for the ACS labor force questions is also a full calendar week, but it is the week prior to the week when the respondent answers the questions rather than a specific week of the month. This week is not the same for all respondents, and, in fact, can vary over all the weeks in a year. To the extent that the labor force characteristics of people in the ACS in their reference week differ from their characteristics in the week containing the 12th, the comparability of ACS and CPS estimates is affected.


In summary, the CPS, ACS, and the LAUS program are complementary rather than competing sources of employment and unemployment data. The CPS has a long history of providing consistent labor force estimates for regular, uniform, time periods. It employs trained field staff and detailed questions. The LAUS program, being pegged to the official concepts and measures of the CPS, benefits from its strengths, and provides timely monthly estimation for numerous substate areas. The strength of the ACS is its coverage and precision of measurement for smaller geographic areas, features neither the CPS nor the LAUS program can replicate, and its wealth of socioeconomic characteristics that can be reliably associated with its labor force measures.


See the Fact Sheet: “Differences between the Labor Force Estimates from the American Community Survey and the Current Population Survey/Local Area Unemployment Statistics Program” for more information on differences among those programs.


For more information on this topic, see the following web sites:


http://www.bls.gov/home.htm

    

http://www.bls.census.gov/cps/cpsmain.htm


http://www.bls.gov/lau/home.htm (particularly the publication, Differences Between Decennial

Census and LAUS Data, at http://www.bls.gov/lau/home.htm#publications )



The ACS and the Census 2000 long form


The ACS uses the same employment status concepts as those used in Census 2000, discussed below. The ACS data, however, are annual averages, whereas the census estimates relate to the period of time when the census was conducted – primarily from March to June 2000.



The CPS and the Census 2000 long form


Employment and unemployment estimates from Census 2000 will, in general, differ from the official labor force data collected in the CPS and released by the BLS, because the design and collection methodology of the census and the CPS meet different purposes.


Census 2000 was designed to collect general information about the labor force for very small geographic areas on a one-time basis. It was primarily a mail-out/mail-back data collection that asked fewer and less detailed questions than the CPS on employment and unemployment. For example, the CPS asks a more detailed series of questions about whether a person is "actively looking for work" than could be asked on a mail-out/mail-back form being answered by people on their own.


At the national level, Census 2000 estimates of employment were below, and estimates of unemployment above, the corresponding CPS estimates. Subnational estimates from the two sources may exhibit even wider relative differences. A known problem in Census 2000 increased the number of unemployed people for some places with relatively large numbers living in civilian non-institutional group quarters, such as college dormitories, worker dormitories, and group homes. This problem may have affected comparisons of labor force data for higher levels of geography. 3


For more information on this topic, see the following web sites:


http://www.census.gov/hhes/www/laborfor/final2_b8_nov6.pdf


http://www.census.gov/pred/www/rpts/B.7%20Final%20Report.pdf (this site presents the text for Census 2000 Evaluation Study B.7, Accuracy of Data for Employment Status as Measured by the CPS-Census 2000 Match )


http://www.census.gov/pred/www/rpts/B7_All%20tables_May4.xls ( this site presents the tables for the above-mentioned Census 2000 Evaluation Study B.7)



Survey of Income and Program Participation (SIPP)


The main objective of the SIPP is to provide accurate and comprehensive national information about the income and program participation of individuals and households, and about the principal determinants of these characteristics. Although the SIPP design allows for both longitudinal and cross-sectional data analysis, SIPP is meant primarily to support longitudinal studies. SIPP’s longitudinal features allow analysts to understand the dynamics of economic well-being and how such dynamics relate to the characteristics of the population. Examples are changes in income, spells without health insurance, movements into and out of poverty, labor force turnover, spells of unemployment, and associated events such as residential mobility and changes in family structure.


SIPP labor force estimates will generally differ from those of the CPS and the ACS because of differences in their labor force definitions, and in the procedures and designs of the surveys. In SIPP the labor force estimates typically relate to the entire 4-month reference period or to an entire month, but in the ACS and CPS the estimates are based on activities occurring in the reference week. The recall period in SIPP extends up to 4 months, but is only 1 week in CPS and the ACS. The labor force questions used in SIPP are neither identical to those used in the CPS or in the ACS, nor as extensive and probing as those in the CPS. Because of these procedural and design differences, SIPP, CPS, and ACS estimates will likely be different.


For more information on this topic, see the following web site:


http://www.bls.census.gov/sipp/



Employer or Establishment Data


Since employment data from the CPS and the ACS are obtained from respondents in households, they differ from statistics based on reports from individual businesses, farm enterprises, and certain government programs. People employed at more than one job are counted only once in the CPS and the ACS. In statistics based on reports from business and farm establishments, people who work for more than one establishment may be counted more than once. Moreover, some establishment tabulations may exclude private household workers, unpaid family workers, and self-employed people, but may include workers less than 16 years of age, all differences from the CPS and the ACS labor force definition.


An additional difference in the data arises from the fact that people who had a job but were not at work (for example, on vacation) are included with the employed in the CPS and the ACS statistics, whereas these people are excluded from employment figures based on establishment payroll reports if they did not receive pay for their absences. Furthermore, the employment status data in household survey tabulations include people on the basis of place of residence regardless of where they work, whereas establishment data report people at their place of work regardless of where they live. This latter consideration is particularly significant when comparing data for workers who commute between areas and is likely to be more important the smaller the geographic area.


For more information on this topic, see the following web sites:


http://www.bls.gov/home.htm (in particular, see http://www.bls.gov/dolfaq/bls_ques24.htm )

 

http://www.bls.gov/ces/home.htm



Unemployment Insurance Data


For several reasons, the unemployment figures from household surveys are not comparable with published figures on unemployment insurance claims. For example, figures on unemployment insurance claims exclude people who have exhausted their benefit rights, new workers who have not earned rights to unemployment insurance, and people losing jobs not covered by unemployment insurance systems (including some workers in agriculture, domestic services, and religious organizations, as well as self-employed and unpaid family workers). In addition, the qualifications for drawing unemployment insurance differ from the definition of unemployment used in the CPS or the ACS. People working only a few hours during the week and people with a job but not at work sometimes receive unemployment insurance benefits based on a previous job, but are classified as "employed" in the CPS and the ACS. Differences between estimates of the geographical distribution of unemployment arise because the place where claims are filed may not necessarily be the same as the place of residence of the unemployed worker.


For more information on this topic, see the following web sites:


http://www.bls.gov/home.htm


http://www.bls.gov/ces/home.htm


http://workforcesecurity.doleta.gov/unemploy/claims.asp



SUMMARY


The nation’s three major sources of employment and unemployment estimates – household surveys, establishment surveys, and unemployment insurance records – serve different, albeit complementary, purposes. Estimates from the various surveys or programs will almost never match (unless explicitly controlled), because of differences in such things as questionnaire content, data collection methodology, reference-period specifications, and editing procedures, as well as the coverage and rules of the unemployment insurance program. Understanding the distinctions between the three sources is the first step in choosing the appropriate data, to be followed by an appreciation for the strengths and weaknesses of the surveys and programs within a particular source.


The chart below presents recommendations for the appropriate household survey to use to obtain employment and unemployment estimates, at various geographic levels.







Geographic Level

Cross-Sectional Estimates





Longitudinal Estimates

Monthly Measurement of Employment Status

Annual Measurement of Employment Status

Basic Status

Status Classified by Socioeconomic Characteristic

Basic Status

Status Classified by Socioeconomic Characteristic*


2000-2004 (American Community Survey testing phase)

United States

CPS

CPS

CPS

CPS or ACS

SIPP

State

LAUS

CPS (selected states)

CPS

CPS or ACS

SIPP (selected states)

Substate Area of 250,000 or more

LAUS

Not available

CPS or LAUS

CPS or ACS

Not available

Substate Area of less than 250,000

LAUS if available

Not available

LAUS if available

Not available

Not available


2005 and beyond (expected American Community Survey full implementation)

United States

CPS

CPS

CPS

CPS or ACS

SIPP

State

CPS

CPS (selected states)

CPS

CPS or ACS

SIPP (selected states)

Substate Area of 65,000 or more

LAUS if available

Not available

CPS or LAUS if available, otherwise ACS

CPS or ACS

Not available

Substate Area of 20,000 to 65,000

LAUS if available

Not available

LAUS if available, otherwise ACS

ACS

Not available

Substate Area of less than 20,000

LAUS if available

Not available

LAUS if available, otherwise ACS

ACS

Not available

* The CPS and the ACS each provide socioeconomic characteristics that the other lacks entirely or to the same level of detail; the ACS, having a larger sample size, can support more finely detailed tabulations than the CPS. The annual CPS data are available only for states and large metropolitan areas.

 

CPS = Current Population Survey

ACS = American Community Survey

LAUS = Local Area Unemployment Statistics Program

SIPP = Survey of Income and Program Participation



Contact the Demographic Call Center Staff at 301-763-2422 or 1-866-758-1060 (toll free) or visit ask.census.gov for further information on Labor Force Data.

Source: U.S. Census Bureau, Housing and Household Economic Statistics Division
Last Revised: October 31, 2011