Local Area Unemployment Statistics Program

1

Local Area Unemployment
Statistics Program:
Introduction and Overview

Introduction
he Local Area Unemployment Statistics (LAUS) program is a
Federal/State cooperative program which produces monthly employment
and unemployment estimates for approximately 7,300 geographic areas.
The areas include Census Regions, Census Division, all States, the District of
Columbia, Puerto Rico, labor market areas (LMAs), counties, cities with a
population of 25,000 or more, and all cities and towns in New England, regardless
of population. These estimates, which are produced by State employment
security agencies, are key indicators of local economic conditions. They are used
by State and local governments for planning and budgetary purposes and as
determinants of the need for local employment and training services and
programs. LAUS estimates are also used to determine the eligibility of an area
for preferential treatment or benefits under various Federal assistance programs.

T

The underlying concepts and definitions of all labor force data developed from the
LAUS program are consistent with those of the Current Population Survey (CPS).
Monthly estimates for all States, the District of Columbia, New York City, Los
Angeles-Long Beach-Glendale, and the respective balances of New York and
California, are produced using estimating equations based on time-series and
regression techniques. These “signal plus noise” models combine current and
historical data from the CPS, the Current Employment Statistics (CES) program,
and State unemployment insurance (UI) systems. Models are also utilized for five
additional substate areas and their respective State balances. The areas are: the
Chicago-Naperville-Arlington Heights, IL metropolitan division; the ClevelandElyria, OH metropolitan area; the Detroit-Warren-Dearborn, MI metropolitan
area; the Miami-Miami Beach-Kendall, FL metropolitan division; and the SeattleLAUS Program Manual 1-1

Bellevue-Everett, WA metropolitan division. Area and balance-of-State models
are controlled directly to their respective State totals.
Monthly estimates for Puerto Rico are produced from a survey modeled on the
CPS survey. This survey is conducted by Puerto Rico and provided to BLS by the
Puerto Rican Bureau of Employment Security.
Estimates for substate areas (other than the substate modeled areas noted above)
are produced using a standard methodology called the “Handbook” method and
disaggregation techniques. The Handbook method uses data from several
sources, including the CPS, CES, State UI systems, and the American Community
Survey (ACS), to create estimates which are then adjusted to the State CPS-based
measures of employment and unemployment. Handbook estimates are developed
at the county level for all States except in New England, where they are
developed at the minor civil division (MCD) level. MCDs represent all cities and
towns in New England regardless of their population. Multi-county (and multiMCDs) metropolitan areas, metropolitan divisions, and small labor market areas
are the summation of the respective component counties (or MCDs). Estimates
for areas below the county level are prepared for all cities with a population of
25,000 or more, using disaggregation techniques based on 5-year ACS population
estimates and current UI statistics.

LAUS Program Manual 1-2

History
Since the late 1940s, subnational estimates of employment and unemployment
have been developed by States under the guidance of the Federal government.
These estimates were initially developed in response to a need to quantify labor
dislocations during World War II. To meet this need, the War Manpower
Commission developed an estimation program to supply figures on local area
labor and material shortages. With the end of the war, the Labor Department’s
Bureau of Employment Security (BES), now the Employment and Training
Administration (ETA), took over responsibility for manpower programs.
As the need for more detailed statistics increased, there was also a need for more
conformity in estimation in the individual States. In 1950, BES introduced
guidelines on estimation, entitled Techniques for Estimating Employment, and
distributed them to the States. A decade later, revised and updated techniques
were republished in the Handbook on Estimating Unemployment. This was a 70step method of estimating procedures for producing unemployment data for the
State and for labor market areas. The Handbook method used a series of
“building blocks,” including establishment employment and unemployment
insurance data, to produce unemployment rates equivalent to the CPS but without
the high cost of a household survey. As early as 1961, local area unemployment
statistics were used to distribute federal funds to local areas under such programs
as the Area Redevelopment Act.
In 1962, the President’s Committee to Appraise Employment and Unemployment
Statistics (the Gordon Committee) criticized the validity of the Handbook method.
This was followed by a series of independent studies comparing the Handbook
estimates to those from the Census or from the CPS. They reported the existence
of biases and inaccuracies in the Handbook procedures. In 1971, the General
Accounting Office, after a year-long audit of two States’ unemployment
estimating programs, came to the same conclusions, and also found that States
were independently introducing their own changes into the Handbook Method.
The GAO recommended that the States’ procedures be reviewed and monitored in
order to reestablish methodological conformity, in that any State change which
improved the accuracy and comparability of the statistics be integrated into the
methodology, and that “high priority” be given to a general improvement in the
estimating methods.
In the early 1970’s, the Bureau of Labor Statistics (BLS) was publishing CPSbased labor force statistics for selected States and large areas while BES was
publishing Handbook-based statistics for all States and areas. Shortly after the
GAO report was issued, the Office of Management and Budget (OMB), as part of
its review of statistical programs in the Department of Labor, determined that
general purpose statistics should be the responsibility of BLS. In November 1972,
the responsibility for local area unemployment statistics was transferred to BLS.
Therefore, beginning in 1973, BLS (with the cooperation of all States) published
monthly labor force data for all States and labor market areas, based on the
LAUS Program Manual 1-3

Handbook procedures. One year later, BLS introduced the first major revisions to
the program. The revisions had a two-fold purpose: to introduce more conformity
between LAUS and CPS data, and to achieve a greater level of consistency of
procedures among the States.
The most important of the methodological changes introduced by BLS in
November 1973 was the direct use of CPS data. At that time, the CPS was a
nationally-based sample. In order to identify usable State CPS data, a reliability
criterion was established which required that State samples be sufficiently large to
estimate the unemployment level with a coefficient of variation (CV) of no more
than 10 percent at one standard error when the unemployment rate is 6 percent.
Applying this standard resulted in the identification of 19 States and 30
metropolitan areas for which CPS data could be used directly as the annual
average benchmark for 1970-73. During 1974, the Census Bureau revised the
procedure used to weight the State sample data to reflect the universe, which
resulted in a lower estimated variance. Thus, 8 more States were able to be
benchmarked to the CPS. In 1975, BLS contracted with the Census Bureau to
expand the sample by 9,000 households in the 23 remaining States and the
District of Columbia, so that all States were able to be benchmarked to annual
average CPS estimates in 1976.
In 1978, BLS broadened the reliability criterion for the use of CPS data by also
considering monthly data, within the context of a budget proposal to expand the
CPS to yield monthly employment and unemployment data for all States by June
1981. Under the expanded criterion, which specified a 10-percent CV on monthly
data, monthly CPS levels were used directly for 10 States, 2 areas, and their
respective balance-of-State areas. The use of annual average CPS data for the
other 28 metropolitan areas was discontinued at that time, so that all substate
areas not meeting the monthly reliability criterion would be treated the same.
Ultimately, the budget proposal, which initiated the direct use of monthly State
CPS data, was rejected as too costly.
In addition to the 1975-76 increase to the CPS to obtain reliable annual average
data for all States, in 1980, 9,000 households were added to improve the
reliability in the 40 nondirect-use States. A final sample increase of 6,000 was
implemented in 1981 to improve the reliability of data in 30 specific metropolitan
areas, 10 of their central cities, and the respective balance-of-State areas. In 1982,
however, because of the Federal budget cut, the 1981 supplement and one-half of
the 1980 supplement were eliminated.
Another part of the improvement commitment supported by the budget
supplements was a $2.5 million effort to standardize and improve the
unemployment insurance data, which provide the only current unemployment
measure for all substate areas. Funding for this initiative was provided to
BLS in 1975-76, and used through contracts with States to correct and
augment unemployment insurance statistics to make them more appropriate
for use in LAUS estimation. Inconsistencies within and among States were
eliminated, quality control measures were instituted, and manual tallies were
LAUS Program Manual 1-4

replaced with computer-generated tabulations. Through such improvements
as the use of place-of-residence of the claimant, the CPS reference week, and
the elimination of claimants who had earnings due to employment, closer
adherence to CPS concepts was achieved. The resultant improved
unemployment insurance data were implemented in LAUS estimation in
1978.
In July 1985, the CPS redesign based on the incorporation of 1980 Census data
was fully implemented. A key part of the redesign involved a change in the
sample structure of the CPS from a national-based one to a State-based stratified
sample. Based on the redesign and sample restructuring, the reliability of the CPS
data at the State level was improved such that the monthly and annual CVs for
direct use of CPS data were reduced to 8 percent.
In 1986, updated inputs to the Handbook based on 1980 Census data and a
number of important methodological improvements in component groups of the
Handbook employed and unemployed not covered by unemployment insurance
were implemented.
Also in 1986, efforts to utilize econometric techniques to estimate monthly State
employment and unemployment were strengthened. The earliest BLS attempts to
explore regression methods go back to the late 1970’s. In addition to internal
work, BLS contracted with Mathematica Policy Research, Inc., which conducted
extensive State and area research using time-series and cross-sectional models.
Their final report was delivered in 1981.
Internal model research continued through the 1980’s, resulting in the
identification of variable coefficient models as a possible substitute for the
nondirect-use State method used at that time—Handbook estimation adjusted to a
six-month moving average CPS. In 1986, the State Research Group was
established with participation of selected State research directors to facilitate the
evaluation of model-based estimates and to ensure adequate communication of
State needs. In 1987, a subsequent group of State research directors was
established—the Regression Implementation Committee—to further evaluate the
model approach during the one-year period of dual estimation in 1988. In
addition to internal and State review of the model-based estimating method,
Professor Art Dempster of Harvard University participated in the evaluation
effort. The result of these efforts was the implementation of variable coefficient
models in the nondirect-use States in January 1989.
Following the incorporation of the first generation of State econometric models,
model research continued. Since seasonally-adjusted estimates were available for
the direct-use States, efforts were focused on the seasonal adjustment of the
model-based State estimates. A BLS workgroup was established in 1989 to
evaluate the appropriateness of seasonally adjusting the model estimates using the
X-11 ARIMA software used for the CPS. The workgroup’s positive report in the
fall of 1991 led to the introduction of monthly seasonally-adjusted nondirect-use
State estimates in 1992.
LAUS Program Manual 1-5

Also in the early 1990’s, a major effort was undertaken to improve the model
specifications. Research was conducted to explicitly account for important
characteristics of the CPS sample design. This led to better control of the effects
of sampling error on the model estimates. In addition, a more flexible modeling
of State-specific seasonal and trend effects was identified. The resultant secondgeneration models were referred to as “signal-plus-noise” models. These
modeling results were provided to States for comment in early 1993, and were
implemented in 1994.
In January 1996, the Bureau reduced the number of households in the CPS, to
accommodate lower funding levels for the labor force program. One result was
that the sample was no longer sufficient to provide monthly data directly for the
11 large States, New York City, and the Los Angeles Metropolitan Area.
Monthly estimates for these States and areas are now produced based on the timeseries modeling methodology used for the other 39 States and the District of
Columbia. Also in January 1996, the LAUS substate estimation process was
streamlined and input options were eliminated to accommodate the reduction of
resources for the LAUS program.
In January 2005, a major program redesign was implemented. Work on the
Redesign began in Fiscal Year 2001, with a budget initiative to enhance the
quality and quantity of LAUS program statistics. Major LAUS Redesign
components included improvements to the method of State and large area
estimation including real-time benchmarking, extending the model-based
estimation methodology to additional substate areas, improving the methods used
in all other areas through better techniques and input data, and updating the
geography with 2000 Census-based areas.
The 2005 LAUS Redesign introduced the third generation of LAUS models. The
new generation models implemented direct model-based seasonal adjustment with
reliability measures and improved the benchmarking procedure by incorporating
real-time monthly benchmarking. At the same time, 6 area models were
introduced along with corresponding balance-of-State models.
Real-time benchmarking addressed a number of concerns with the prior
generation of LAUS models. It reduced annual revisions by incorporating the
CPS benchmark on a current basis. It eliminated prior model biases and
benchmarking issues. It ensured that national events and shocks to the economy
were reflected in State estimates as they occurred. It also eliminated the
discrepancy between the sum-of-States estimates and the national not-seasonallyadjusted totals. Measures of error on the seasonally-adjusted and not-seasonallyadjusted estimates and the over-the-month change were introduced for Division,
State, and area model estimates.
The LAUS Redesign also included projects to improve methodology, update
geography and decennial census inputs, and improve the quality of inputs to the
estimates. The Redesign resulted in significant improvements to the accuracy of
LAUS Program Manual 1-6

the LAUS labor force estimates and has enhanced the ability to analyze labor
market behavior. Methodological changes included improvements to substate
unemployment estimation that addressed long-standing inadequacies with the
previous method and an innovative approach to adjusting place-of-work
employment to place-of-residence that more accurately reflects complexities of
commuting. In addition, all LAUS areas were revised for the latest OMB and
BLS geographic definitions.
In January 2010, BLS introduced additional improvements to the LAUS models
with the implementation of smoothed seasonally adjusted (SSA) estimates. The
SSA estimates incorporate a long-run trend smoothing procedure, resulting in
estimates that are less volatile than those previously produced by the LAUS
estimation methodology. The use of SSA methodology is effective in reducing
the number of spurious turning points in current estimates. More importantly,
SSA estimation can reduce revisions in historical estimates and remove the
potential disconnection between historically benchmarked and current estimates.
The most current program redesign was implemented in 2015. This included the
introduction of fourth generation models, the replacement of decennial census
long-form data with ACS data, improved smoothed seasonal adjustment, and
updated procedures for developing substate area estimation.
The fourth generation models introduced major improvements including modelbased benchmarking where the models directly produce estimates that
automatically sum to Census Division controls and thus eliminated the need for
external pro-rata factors previously used to benchmark state estimates to Census
Divisions. This approach provides greater flexibility, smoother monthly
adjustment factors, and improved reliability measures. These models also allow
for the additivity of outlier effects. Outlier estimates are separated from the
benchmarking process, resulting in outliers being specific to where they occurred,
rather than spread across all states within a Census Division. The new model
structure uses CES and UI trend estimates as regressor values to explain trend
variation in the CPS, which produces similar results to the bivariate models, but
with a major reduction in computing time and allows for more flexibility with
model development over the long term.
The LAUS program had been reliant on decennial Census long-form data as the
basis for adjusting establishment-based employment estimates to residency-based
employment estimates, for estimating certain employment and unemployment
components in the Handbook methodology, and for disaggregating or
apportioning labor market area estimates to smaller areas. With the
discontinuation of the long form for the 2010 Census, these inputs were replaced
by 5-year estimates from the ACS. These data are the most statistically reliable of
the ACS estimates and cover all LAUS geography.

LAUS Program Manual 1-7

LAUS Time Line
Year

Historical Developments Related to LAUS

1933

Wagner-Peyser Act created Employment Service for
registering the unemployed

1935

Social Security Act created Unemployment Insurance
System

1937

Works Projects Administration began collecting
household-based labor force data

19391945

War Manpower Commission developed program on
local area labor and material shortages

1943

Responsibility for conducting household survey
transferred to Bureau of the Census

1948

Monthly Report on the Labor Force renamed Current
Population Survey (CPS)

1950

BES (now ETA) published the manual “Techniques for
Estimating Unemployment”

1959

Responsibility for analyzing and publishing CPS data
given to BLS; Census continues to conduct survey

1960

Manual for estimating area unemployment revised by
BES, title changed to “Handbook on Estimating
Unemployment” (70-Step Method)

1961

Area Redevelopment Act passed

1962

President’s Committee to Appraise Employment &
Unemployment Statistics (Gordon Committee) issued
final report

1965

Public Works and Economic Development Assistance
Act (PWEDA) passed

1972

Responsibility for LAUS program transferred to BLS

1973

Comprehensive Employment and Training Act (CETA)
passed

1975

CPS sample expansion; CPS benchmarking extended to
27 States

1975

First round of UI Database Survey conducted by BLS

1976

CPS benchmarking extended to all States.

1976

Public Works Employment Act (PWEA) passed

LAUS Program Manual 1-8

LAUS Time Line (Continued)
Year

Historical Developments Related to LAUS

1976

National Commission on Employment and
Unemployment Statistics (Levitan) and National
Commission on Unemployment Compensation established

1978

Direct use of monthly CPS estimates for limited
number of States and areas introduced

1978

First UI database improvements incorporated into the
Handbook estimates

1979

Levitan Commission issued recommendations

1982

Job Training Partnership Act (JTPA) replaced CETA

1983

Second round of UI Database Survey conducted by
BLS

1985

Updated State-based CPS sample based on 1980
Census introduced

1986

Major revisions to Handbook methodology
incorporated

1989

Variable coefficient model estimates incorporated for
nondirect-use States

1992

Seasonal adjustment of model-based estimates
introduced.

1994
1996

Second generation of LAUS models introduced; 1990
Census data incorporated into LAUS; new CPS
questionnaire and data collection method implemented
Direct-use States adopt model based estimation
method; Handbook method streamlined to 13 steps

2005

Third generation of LAUS models introduced, bringing
real-time benchmarking and model-based seasonal
adjustment to the methodology; improvements to
Handbook methodology include dynamic ratio
adjustment for place-of-work employment and updated
new and reentrant unemployed estimation; 2000
Census data incorporated into LAUS; and redesign of
the State estimating system.

2010

Smoothed-seasonally-adjusted methodology
implemented to develop official seasonally-adjusted
estimates
LAUS Program Manual 1-9

LAUS Time Line (Continued)
Year

2015

Historical Developments Related to LAUS
Fourth generation of LAUS models improved State
time-series estimation with the introduction of modelbased benchmarking that accounts for errors in the
estimates, additivity of outlier effects that allocates
level shifts to the appropriate State, a more efficient
model structure that reduces processing time, and
enhanced smoothed seasonal adjustment procedures.
Data from the ACS replaced decennial census inputs
and updated procedures for developing substate
estimates were also introduced.

LAUS Program Manual 1-10

Data Sources
LAUS estimates are designed to reflect the labor force concepts embodied in the
CPS and, thus, are conceptually comparable to each other.
LAUS estimates are based on data from a number of different sources. Primary
source data for the creation of employment and unemployment estimates include
the CPS; the State UI systems; the CES program; the QCEW; Census’ Population
Estimates Program (PEP); and the American Community Survey. Each of these
inputs to LAUS estimation is described in detail in the following four chapters. A
brief summary of each data source is provided below.

The Current Population Survey
The CPS is a monthly sample survey of households, conducted by the Bureau of
the Census under contract to the BLS. It provides statistics on the labor force
status of the civilian noninstitutional population 16 years of age and over. CPS
data are collected each month from a probability sample of approximately 60,000
occupied households and yield estimates of demographic, social, and economic
characteristics of the population.
The BLS has responsibility for analyzing and publishing monthly employment
and unemployment estimates for the Nation. CPS data are valuable inputs into
LAUS monthly estimation due to their regular availability, comparability across
States, and measurable statistical error. (See Chapter 2 for more details on the
CPS.)

Unemployment Insurance Systems
Under the UI system, an employer must pay a tax for each employee covered by
the State law. Coverage includes the State UI program and the Federal Civilian
Employment program. This tax is, in effect, an insurance premium paid to
provide for possible unemployment benefits. When any employee in a covered
job becomes unemployed, he/she may file an initial claim to receive
unemployment insurance benefits. A determination of whether the individual is
covered by the State unemployment insurance system and, if so, how much in
benefits is due, will then be made by the State. A qualifying claimant will receive
weekly compensation until the maximum benefit amount is exhausted or until the
person returns to work, whichever is earlier.
The UI administrative statistics created in this process are useful for LAUS
estimation because they are current and are generally available for a great many
geographic areas. (See Chapter 3 for more details on the UI system.)
LAUS Program Manual 1-11

Current Employment Statistics and Quarterly Census of
Employment and Wages
Both the CES and QCEW programs are Federal/State cooperative ventures which
obtain employment data from employers. The CES is a voluntary sample survey
of establishments covered by State and Federal UI laws. It is designed to produce
monthly estimates of employment, hours, and earnings for the Nation, all States,
and most major metropolitan areas.
The QCEW data series is a universe of monthly employment and quarterly wage
information by industry, county, and State. Completion of a quarterly
contribution report, which is basis for the QCEW, is mandatory in industries
covered by Federal and State UI laws.
Data obtained through these two programs are used in LAUS employment
estimation. (See Chapter 4 for a detailed discussion of the CES and QCEW
programs.)

American Community Survey
The ACS is a household survey developed by the Census Bureau to replace the
long form of the decennial census program. The ACS is a large survey collected
throughout the year using mailed and internet questionnaires, telephone
interviews, and visits from Census Bureau field representatives to about 3 million
household addresses annually. Starting in 2005, the ACS produced social,
housing, and economic characteristic data for demographic groups in areas with
populations of 65,000 or more. The ACS also accumulates samples over 3-year
(discontinued with the 2011-2013 dataset) and 5-year intervals to produce
estimates for smaller geographic areas, including census tracts and block groups.
Five-year estimates from the ACS are used in LAUS estimation for adjusting
establishment-based employment estimates to residency-based employment
estimates, for estimating certain employment and unemployment components in
the Handbook methodology, and for disaggregating or apportioning labor market
area estimates to smaller areas. (See Chapter 5 for additional details on the ACS.)

LAUS Program Manual 1-12

Summary of Estimation Methods
Monthly estimates of employment and unemployment are prepared for
approximately 7,500 geographic areas, which include all States, labor market
areas, counties, cities with a population of 25,000 or more, and all cities and
towns in New England, regardless of population. At each level of geographic
detail, the estimation method used depends on the most current data sources
available.
Statewide Estimates
A tiered approach to estimation is used for statewide estimates. First, modelbased estimates are developed for the nine Census divisions that geographically
exhaust the nation using univariate signal-plus-noise models. The Division
models are similar to the state models, but do not use unemployment insurance
claims or payroll employment as input variables. The division estimates are
benchmarked to the national levels of employment and unemployment on a
monthly basis. The benchmarked division model estimate is then used as the
benchmark for the states within the division.
Monthly labor force estimates for all States, the District of Columbia, the Los
Angeles-Long Beach-Glendale metropolitan division, New York City, and the
respective balances of California and New York are based on dynamic time-series
regression models that utilize data from the CPS, UI systems, and the CES survey.
Both seasonally-adjusted and unadjusted estimates are produced each month.
The model methodology is also utilized for five additional substate areas and their
respective balances of States. The areas are: the Chicago-Naperville-Arlington
Heights, IL metropolitan division; the Cleveland-Elyria, OH metropolitan area;
the Detroit-Warren-Dearborn, MI metropolitan area; the Miami-Miami BeachKendall, FL metropolitan division; and the Seattle-Bellevue-Everett, WA
metropolitan division. These models are covariate, like the state and area models
mentioned above, in that they use UI and CES inputs. However, the substate area
and the balance of state estimates produced by these models are benchmarked to
the statewide control totals of employment and unemployment. (See Chapter 6.)
Substate Labor Market Estimates
States are divided into Labor Market Areas (LMAs), which exhaust the
geographic area of the State. Independent estimates are produced for all counties
except in New England, where they are produced for all New England City and
Town Areas (NECTAs) using a standard procedure known as the “Handbook”
method. The Handbook method is an effort to estimate employment and
unemployment for an area using available information, comparable to what would
be produced by a random sample of households in the area, without the expense
of a large labor force survey like the CPS. The county and NECTA level
Handbook estimates are controlled to the LAUS Statewide estimates to create the
official estimates. County and NECTA estimates are aggregated to LMA level
LAUS Program Manual 1-13

estimates. LAUS estimates for cities with populations over 25,000 are derived by
a disaggregation technique using population estimates and UI statistics, or data
from the ACS. (See Chapters 7, 8, and 9 for further details.) At the end of the
year, State and substate areas are revised and benchmarked to reflect updated,
revised data. (See Chapter 10.)

LAUS Program Manual 1-14

LAUS Estimation Techniques

Area

Estimation Method

Nine Census Divisions

Signal-plus-noise univariate
regression model

50 States

Signal-plus-noise covariate
regression model

District of Columbia

Signal-plus-noise covariate
regression model

New York City, Balance of NY
State

Signal-plus-noise covariate
regression model

Los Angeles, Balance of
California

Signal-plus-noise covariate
regression model

Chicago, Cleveland, Detroit,
Miami, Seattle and balances of
Illinois, Ohio, Michigan,
Florida, and Washington

Signal-plus-noise covariate
regression model

Remaining Labor Market
Areas (LMAs) and counties

Handbook, Additivity

Cities over 25,000 population

Disaggregation

LAUS Program Manual 1-15

Publication and Administrative Uses of LAUS Estimates
LAUS was given responsibility to develop and publish the most current national,
State, and local labor force and unemployment data by the OMB in Statistical
Policy Directive No. 11, “Standard Data Source for Statistical Estimates of Labor
Force and Unemployment.” This directive also requires the use of BLS-developed
LAUS estimates by federal executive departments, agencies, and establishments in
allocations of federal resources and eligibility determinations. The complete text
of this Directive is provided at the end of this chapter.

Publication of LAUS Estimates
Each month, labor force, employment, unemployment, and
unemployment rate estimates for all LAUS areas are published
by BLS. Data from the LAUS program are made available to
users in a variety of ways.


The monthly “Regional and State Employment and
Unemployment” news release is issued approximately
two weeks after the national release of labor force data.
It presents data for the Census Regions, and Divisions
and all modeled LAUS estimates. The Bureau’s public
database, LABSTAT http://www.bls.gov/lau/), is
also updated with these estimates at that time.



The monthly “Metropolitan Area Employment and
Unemployment” news release is issued about 12
days after the Region and State release and contains
labor force and unemployment estimates for all
metropolitan areas in the nation. These estimates are
also issued in LABSTAT at that time.



Estimates of labor force, employment, and unemployment for micropolitan
areas, small labor market areas, counties, cities with a population of 25,000
or more, and all cities and towns in New England are issued in LABSTAT
at the same time the Metropolitan area data are released.



Annual average employment status data are provided each year in a press
release entitled “State and Regional Unemployment, Annual Averages”,
which is typically issued at the end of February. It presents data on the
population, civilian labor force, employed, unemployed, and
unemployment rate for regions, Divisions, and States.



The annual publication, Geographic Profile of Employment and
Unemployment, provides annual average CPS data for Census Regions and
LAUS Program Manual 1-16

Divisions, the 50 States and the District of Columbia, 54 large metropolitan
areas, 22 metropolitan divisions and 41 principal cities. Data are provided
on the employed and unemployed by selected demographic and economic
characteristics.

Legislative Uses of LAUS Estimates
Each year, LAUS estimates are used to distribute federal funds to States and areas
or make eligibility determinations by a number of federal programs. The
following table, “Administrative Uses of Local Area Unemployment Statistics”,
presents information on the federal programs that utilize LAUS data in allocating
funds to States. Allocation formulas, reference periods, and geographic coverage
information are presented. Total funding for these programs amounted to
$90,341.7 million in Fiscal Year 2013. These programs are described in greater
detail in the section below. The American Recovery and Reinvestment Act of
2009 used LAUS data to allocate an additional $144,350.0 million to States.

LAUS Program Manual 1-17

Department of Labor:
Employment and Training
Administration, Workforce Investment
Act, Title 1, Chapter 5: Adult
Employment and Training Activities
Program Objectives: To provide millions of adult
workers with workforce preparation and career
development services, and help employers find the
skilled workers they need. Activities promote and
facilitate an integrated public workforce system through which a full array of services is
offered. These services are available to workers and employers through the national
network of One-Stop Career Centers. Programs provide high-quality employment and
training services that address the needs of individuals in need of training, retraining, and
skill upgrades. Additionally, investments in adult services are targeted to move workers
in post-secondary educational pipelines and career pathways to prepare more workers to
enter into and advance in good jobs in the high growth and emerging occupations of the
global economy.

Employment and Training Administration, Workforce
Investment Act, Title 1, Chapter 4: Youth Activities
Program Objectives: To increase the number of youth entering employment, postsecondary education, or advanced training; the number of youth attaining a degree or
certificate; and literacy and numeracy gains. Provides tutoring, alternative secondary
school offerings, summer employment opportunities linked to academic and occupational
learning, paid and unpaid work experiences, occupational skill training, leadership
development opportunities, supportive services, mentoring, follow-up services, and
comprehensive guidance and counseling. Targeted to youth, aged 14 through 21, who are
low income and have one or more of the following barriers: deficiency in basic literacy
skills; school dropout; homeless, runaway, or foster child; pregnant or parenting;
offender; or require additional assistance to complete an educational program, or to
secure and hold employment.

Employment and Training Administration, Workforce
Investment Act, Title 1, Chapter 5: Dislocated Worker and
Training Activities
Program Objectives: To provide quality employment and training services to assist
dislocated workers in finding and qualifying for meaningful employment, and to help
employers find the skilled workers they need to compete and succeed in
business. Dislocated workers are individuals who: have been terminated, laid off, or
have received notice of termination or layoff from employment; are eligible for or have
exhausted unemployment insurance; have demonstrated appropriate attachment to the
workforce, but are not eligible for unemployment insurance and are unlikely to return to a
LAUS Program Manual 1-18

previous industry or occupation; have been terminated or laid off as a result of a
permanent closure or substantial layoff; is employed at a facility that will be closed
within 180 days; were self-employed but are now unemployed as a result of general
economic conditions or because of a natural disaster; or are displaced homemakers who
are no longer supported by another family member.

Employment and Training Administration, Employment Service
Grants to States
Program Objectives: To assist persons to secure employment and workforce information
by providing a variety of job search assistance and information services without charge to
job seekers, including persons with disabilities and to employers seeking qualified
individuals to fill job openings. Grants are made to the states to provide an integrated
array of high-quality services that workers, job-seekers, and businesses can access under
one roof in easy-to-reach locations, with many services also offered through self-service
electronic access.
The Disability Employment Initiative awards competitive grants to state WIAadministering entities to improve educational, training, and employment opportunities
and outcomes of youth and adults with disabilities who are unemployed, underemployed,
and receiving Social Security disability benefits.

Employment and Training
Administration, Labor Surplus Areas
Program Objectives: The purpose in classifying labor surplus areas is to put the
Federal Government’s procurement contracts into areas of high unemployment.
Employers located in these labor surplus areas are eligible for preference in bidding on
Federal procurement contracts to direct government funds into areas where people are
in the most severe economic need.

Employment and Training Administration, Federal-State
Extended Unemployment Compensation Program
Program Objectives: Unemployment compensation is designed to provide benefits to
most workers out of work due to no fault of their own for periods between jobs. Most
States pay a maximum of 26 weeks of benefits, though an increasing number of States
pay less than 26 weeks. In periods of very high unemployment in individual States,
benefits are payable for as many as 13 additional weeks, up to a maximum of 39
weeks. These “extended benefits” are funded on a shared basis: approximately half
from State funds and half from Federal sources.

Employment and Training Administration, Youthbuild
Program
Program Objectives: To provide programs for low-income young people ages 16 to 24
working full-time for 6 to 24 months toward their GEDs or high school diplomas while
LAUS Program Manual 1-19

learning job skills by building affordable housing for homeless and low-income people
in their communities. Emphasis is placed on leadership development, community
service, and the creation of a positive mini-community of adults and youth committed
to each other’s success. Students may earn AmeriCorps education awards through their
homebuilding and other community service. At exit, they are placed in college, jobs, or
both.

Employment and Training Administration, Senior Community
Service Employment Program
Program Objectives: A community service and work-based job training program for
older Americans. Authorized by the Older Americans Act, the program provides
training for low-income, unemployed seniors. Participants also have access to
employment assistance through American Job Centers.
SCSEP participants gain work experience in a variety of community service activities
at non-profit and public facilities, including schools, hospitals, day-care centers, and
senior centers. The program provides over 40 million community service hours to
public and non-profit agencies, allowing them to enhance and provide needed services.
Participants work an average of 20 hours a week, and are paid the highest of federal,
state or local minimum wage. This training serves as a bridge to unsubsidized
employment opportunities for participants. Participants must be at least 55,
unemployed, and have a family income of no more than 125% of the federal poverty
level. Enrollment priority is given to veterans and qualified spouses, then to
individuals who are over 65, have a disability, have low literacy skills or limited
English proficiency, reside in a rural area, are homeless or at risk of homelessness,
have low employment prospects, or have failed to find employment after using services
through the American Job Center system.

Veterans Employment and Training Service, Jobs for Veterans
Act of 2002
Program Objectives: Employment and training services are provided to veterans of the
United States Armed Forces through a nationwide network of approximately 3,000
One-Stop Career Centers. Priority of Service is granted for veterans and eligible
spouses in all qualified job training programs including: state and local Workforce
Investment Boards; private, national and/or pilot/demonstration operators of
employment and training programs funded by the Department of Labor; those
programs implemented by states or local service providers based on Federal Block
grants administered by DOL; and any such program or service that is a workforce
development program targeted to specific groups.

LAUS Program Manual 1-20

Department of Agriculture:
Temporary Emergency Food
Assistance Program
Program Objectives: To make funds available to States for
storage and distribution costs incurred by nonprofit eligible
recipient agencies in providing nutrition assistance in
emergency situations and to aid needy people. TEFAP was
created to reduce excess USDA inventories of surplus
commodities in storage, especially dairy products such as
cheese, and to supplement the diets of low-income households at a time of high
unemployment. Each State designates one agency to administer TEFAP. Once USDA
commodities are made available to the States, State officials are responsible for
determining the eligibility of organizations to receive the commodities and for entering
into agreements regarding allocation and distribution. In addition, States are responsible
for determining the types and amounts of each commodity to be made available to
organizations within the State.
The Personal Responsibility and Work Opportunity Reconciliation Act of 1996
provided for the absorption of the Soup Kitchens/Food Banks Program into TEFAP..

Waivers to Supplemental Nutrition Assistance Program (SNAP)
and Time Limits for Able-Bodied Adults without Dependents
(ABAWD)
Program Objectives: The Personal Responsibility and Work Opportunity Reconciliation
Act of 1996 limits receipt of SNAP benefits to 3 months in a 3-year period for ablebodied adults who are not working, participating in a work program for 20 hours or more
each week, or in workfare. States may request a waiver of this provision in areas with an
unemployment rate above 10 percent, or for those residing in an area that has an
insufficient number of jobs to provide employment for individuals.
In addition, waiver of this provision may also occur in recognition of the challenges that
low-skilled workers may face in finding and keeping permanent employment. In some
areas, including parts of rural America, the number of unemployed persons and the
number of job seekers may be far larger than the number of vacant jobs. This may be
especially so for persons with limited skills and minimal work history.

LAUS Program Manual 1-21

Department of Commerce:
Economic Development
Administration, Public Works and
Economic Development Program
Program Objectives: To assist States and local areas
in the development and implementation of strategies
designed to arrest and reverse the problems
associated with long-term economic deterioration.
Grants are provided to help distressed communities
attract new industry, encourage business expansion,
diversify local economies, and generate long-term,
private sector jobs.
Among the types of projects funded are water and sewer facilities primarily serving
industry and commerce; access roads to industrial parks or sites; port improvements; and
business incubator facilities. Proposed projects must be located within an EDAdesignated Redevelopment Area (RA) or Economic Development Center. Projects in
other areas of an EDA-designated Economic Development District are also eligible if
they will directly benefit a RA within the District. Projects must be consistent with an
approved Overall Economic Development Program (OEDP). An applicant may be a
state, political subdivision of a state, Indian tribe, special-purpose unit of government, or
a public or private nonprofit organization or an association representing the RA or part
thereof.

Economic Development Administration, Economic
Adjustment Assistance Program
Program Objectives: The Economic Adjustment Program helps States and local areas
design and implement strategies for facilitating adjustment to changes in their economic
situation that are causing or threaten to cause serious structural damage to the
underlying economic base. Such changes may occur suddenly (Sudden and Severe
Economic Dislocation) or over time (Long-Term Economic Deterioration) and result
from industrial or corporate restructuring, new Federal laws or requirements, reductions
in defense expenditures, depletion of natural resources, or natural disasters.
Strategy grants provide the recipient with the resources to organize and carry out a
planning process resulting in an adjustment strategy tailored to the particular economic
problems and opportunities of the impacted area(s). Implementation grants may be used
to support one or more activities identified in an adjustment strategy approved, though
not necessarily funded, by EDA. Implementation activities may include, but are not
limited to: the creation or expansion of strategically targeted business development and
financing programs including grants for revolving loan funds, infrastructure
improvements, organizational development, and market or industry research and analysis.

LAUS Program Manual 1-22

Department of Defense:
Defense Logistics Agency,
Procurement and Technical
Assistance
Program Objectives: Procurement Technical Assistance
Centers (PTACs) serve as a resource for businesses
pursuing and performing under government contracts,
including contracts with the Department of Defense, other
federal agencies, state and local governments and with
government prime contractors. PTACs are hosted by organizations such as universities
and local chambers of commerce. The training and assistance provided by the PTACs is
usually free of charge. PTAC support to businesses includes registration in systems such
as the System for Award Management, identification of contract opportunities, help in
understanding requirements and in preparing and submitting bids.
The PTACs have a local presence in all 50 states, the District of Columbia, Puerto Rico,
and Guam. They are funded through cost sharing cooperative agreements between the
DLA and eligible participants, including states, local governments, nonprofit
organizations, economic enterprises, and tribal organizations.

Department of Health and Human
Services:
Administration for Children and
Families, Temporary Assistance for
Needy Families Contingency Fund
Program Objectives: Temporary Assistance for Needy
Families (TANF) was established under the Personal
Responsibility and Work Opportunity Reconciliation Act
of 1996 to replace Aid to Families with Dependent
Children (AFDC), Job Opportunities and Basic Services (JOBS), and Emergency
Assistance (EA) programs. In order to receive the new block grants under TANF, States
must submit a State TANF plan outlining how they intend to conduct a program that
provides assistance to needy families with children and provide parents with job
preparation, work, and support services to enable them to leave the program and become
self-sufficient. States must submit plans every two years and may submit amendments to
keep the plan current whenever they wish to make changes in the administration or
operation of the program. In addition to State plans, federally recognized Indian Tribes
and approved Alaskan Native entities are also eligible to submit TANF plans to the
Secretary of Health and Human Services.
LAUS Program Manual 1-23

Administration for Children and Families, Temporary Assistance
for Needy Families–Job Search and Job Readiness Activities
Program Objectives: The time limit on participation in job search and job readiness is
extended to 12 weeks if the state unemployment rate is at least 50 percent greater than the
unemployment rate of the United States or if the state meets the definition of a “needy
state” under Contingency Fund provisions. There are two ways for a state to qualify as a
“needy state:” based on its unemployment rate, or based on increases in its Food Stamp
caseload.

Substance Abuse and Mental Health Services Administration,
Community Mental Health Block Grants
Program Objectives: Makes funds available to provide community mental health
services. Support the grantees in carrying out plans for providing comprehensive
community mental health services. Grantees can be flexible in the use of funds for both
new and unique programs or to supplement their current activities. The program targets:
Adults with serious mental illnesses and children with serious emotional disturbances.
Funds may be distributed to local government entities and non-governmental
organizations. They must ensure that community mental health centers provide such
services as screening, outpatient treatment, emergency mental health services, and day
treatment programs.

Substance Abuse and Mental Health Services Administration,
Prevention and Treatment of Substance Abuse Block Grants
Program Objectives: Grantees use the block grant programs for prevention, treatment,
recovery support, and other services to supplement Medicaid, Medicare, and private
insurance services. Specifically, block grant recipients use the awards for the following
purposes: Funding priority treatment and support services for individuals without
insurance or for whom coverage is terminated for short periods of time; fund those
priority treatment and support services that demonstrate success in improving outcomes
and/or supporting recovery that are not covered by Medicaid, Medicare, or private
insurance; fund primary prevention by providing universal, selective, and indicated
prevention activities and services for persons not identified as needing treatment; and
collect performance and outcome data to determine the ongoing effectiveness of
behavioral health promotion, treatment, and recovery support services.

LAUS Program Manual 1-24

Department of Homeland Security:
Federal Emergency Management Agency,
Emergency Food and Shelter Program
Program Objectives: To help meet the needs of hungry and
homeless people in the U.S. and its territories by allocating
Federal funds to the neediest areas, ensuring quick response,
fostering public and private cooperation, ensuring local
decision-making, and maintaining minimal reporting. The
program began in 1983 to help meet the needs of hungry and
homeless people throughout the United States and its territories by allocating federal
funds for the provision of food and shelter to those impacted by natural disasters or
emergencies. The program is governed by a national board composed of representatives
of the American Red Cross; Catholic Charities, USA; Council of Jewish Federations; the
National Council of the Churches of Christ in the USA; the Salvation Army; and the
United Way of America. The national board awards funds to jurisdictions based upon a
formula. Once an award is made, local boards decide which agencies are to receive funds,
and then those agencies are paid directly by the national board.
Funds are used to provide the following, as determined by the local board in funded
jurisdictions: (1) food, in the form of served meals or groceries; (2) lodging in a mass
shelter or hotel; (3) one month’s rent or mortgage payment; (4) one month’s utility bill;
(5) minimal repairs to allow a mass feeding or sheltering facility to function during the
program year; and (6) equipment necessary to feed or shelter people, up to $300 per item.

Federal Emergency Management Agency, National Pre-Disaster
Mitigation Program
Program Objectives: Provides funds for hazard mitigation planning and projects on an
annual basis. The PDM program was put in place to reduce overall risk to people and
structures, while at the same time, also reducing reliance on federal funding if an actual
disaster were to occur. Grants are made to states, territories, and tribal governments and
may be distributed to individual homeowners and businesses through state agencies and
local governments.

U.S Citizenship and Immigration Services, Employment Creation
Visa (EB-5) Program
Program Objectives: To stimulate the U.S. economy through job creation and capital
investment by foreign investors. All EB-5 investors must invest in a new commercial
enterprise and create or preserve at least 10 full-time jobs for qualifying U.S. workers
within two years of the immigrant investor’s admission to the United States. The
investor may create or preserve either direct or indirect jobs. The required minimum
investment is $1 million or $500,000 within a high-unemployment area or rural area in
LAUS Program Manual 1-25

the U.S.

Department of the Treasury:
Riegle Community Development
and Regulatory Improvement Act of
1994, Bank Enterprise Award
Program
Program Objectives: To promote the formation and
expansion of Community Development Financial
Institutions (CDFIs); promote community lending and
investment activities by banks and thrifts; enhance the
liquidity of community lending products; and enhance the capacity of CDFIs, banks and
thrifts to engage in community lending and investment activities. The Bank Enterprise
Award Program is intended to encourage banks and thrifts to invest in and support
community development financial institutions and to increase the lending and services
provided in distressed communities by traditional financial institutions.

Riegle Community Development and Regulatory Improvement
Act of 1994, Community Development Financial Institutions
Fund (CDFI)
Program Objectives: To increase economic opportunity and promote community
development investments for underserved populations and in distressed communities in
the United States.
The CDFI Fund was created for the purpose of promoting economic revitalization and
community development through investment in and assistance to community
development financial institutions (CDFIs). The CDFI Fund was established by the
Riegle Community Development and Regulatory Improvement Act of 1994.

North American Development Bank, Community Adjustment and
Investment Program
Program Objectives: To assist in the creation and/or preservation of private sector jobs
within CAIP eligible communities. The CAIP does this primarily by assisting with the
provision of credit. This can be accomplished with direct assistance or through the
CAIP's partnership with certain loan programs of the Small Business Administration and
the U.S. Department of Agriculture's Business and Industry Loan Guarantee Program.

LAUS Program Manual 1-26

Appalachian Regional Commission:
Area Development Program,
Distressed Counties Grants
Program Objectives: Targets special resources to the most
economically distressed counties in the Region, using a measure of
economic distress based on three economic indicators: three-year
average unemployment rates, per capita market income, and poverty
rates. Besides allocating funding to benefit distressed counties, ARC
has established other policies to reduce economic distress. ARC normally limits its
maximum project funding contribution to 50 percent of costs, but it can increase its
funding share to as much as 80 percent in distressed counties.

Small Business Administration:
Historically Underutilized Business
Zones (HUBZones)
Program Objectives: To encourage economic development
and create jobs in urban and rural communities by providing
contracting preferences to small businesses located in and
hiring employees from historically underutilized business
zones. A firm may be determined to be a qualified
HUBZone small business if it is located in a historically
underutilized business zone, it is owned and controlled by one or more U.S. citizens, and
at least 35 percent of its employees reside in a HUBZone.
Under the program, three types of contracts exist: (1) A competitive contract, in which at
least two qualified small businesses are expected to submit offers, and at least one of
which will be at a fair market price; (2) a sole source contract, and (3) an open
competition award, in which a qualified HUBZone small business receives a price
preference over another non-HUBZone bidder that is not small.

LAUS Program Manual 1-27

Office of Management and Budget Statistical
Policy Directive Number 11
Standard Data Source for Statistical Estimates
of Labor Force and Unemployment
Accurate, consistent, publicly available estimates of the labor force and of
unemployment in the Nation, the States, and local areas are needed for
use in the formulation, implementation, and evaluation of public policy.

1. Source of Data
Federal executive branch departments, agencies, and establishments
(hereinafter Federal executive branch agency) shall use the most current
national, State, or local area labor force or unemployment data published
by the BLS, United States Department of Labor, with respect to all
program purposes, including the determination of eligibility for and/or the
allocation of Federal resources, requiring the use of such data unless
otherwise directed by statute. In order to maintain equity among local
areas, comparable data series are to be used for all program purposes.
Further, unless otherwise required by statute, data adjusted for seasonal
variation shall be used for all program purposes as soon as the BLS shall
have published such data for local areas being examined for the program
purpose then under consideration.
No Federal executive branch agency shall begin or continue collecting or
using State or local area labor force or unemployment data other than that
published by the BLS, without the written approval of the Secretary of
Commerce. This does not preclude the collection of labor force and
unemployment data by the Bureau of the Census, United States
Department of Commerce, for the BLS or in its conduct of a periodic or
other census or statistical survey, and the publication or other distribution
thereof.

2. Data Consistency
With respect to any month, a consistent reference time period shall be
used for all national, State, and local area labor force and unemployment
data. The data for each State and area, to the extent technically feasible,
shall be conceptually consistent with the data for the Nation as a whole
and the State totals shall sum, within a range of acceptable sampling error,
to the national total.

LAUS Program Manual 1-28

3. Data Publication
The BLS, in accordance with the provisions of Directive No. 4, Prompt
Compilation and Release of Statistical Information, shall establish a
monthly release date or dates for all regularly published labor force and
unemployment data and shall provide the release date schedule to the
Office of Federal Statistical Policy and Standards for publication in the
Statistical Reporter.
The monthly publication or publications by the BLS shall contain data for
the Nation as a whole, and for each State and each local area for which the
BLS has agreed to publish data. No agreement between the BLS and other
Federal executive branch agencies shall be used to limit the number or
types of areas for which data are developed and/or published by the BLS.
The data published by area shall at a minimum provide the current
estimates before seasonal adjustment, and as soon as possible, and to the
extent technically feasible, shall also provide the estimate adjusted for
seasonality.

4. Notification of Data Need
Federal executive branch agencies requiring State and local area labor
force or unemployment data shall notify the Commissioner, Bureau of
Labor Statistics, United States Department of Labor, of their need for such
data. The notification shall include information about the purpose for
which the data are needed and the specification(s) (i.e., statistical
reliability, geographic and other) for the data.
Any Federal executive branch agency required by legislation to use labor
force or unemployment data other than that directed by this Directive and
any Federal executive branch agency notified by the Commissioner,
Bureau of Labor Statistics that the needed data cannot be provided
according to specification shall notify the Director, Office of Federal
Statistical Policy and Standards, Department of Commerce of that fact.
The notification shall include identification of the program(s) affected,
legislation implemented by those programs, data specifications, and a
report on consultations with the Bureau of Labor Statistics in respect to
such data.

5. Definitions
a. Labor Force and Unemployment Data. The term labor force and
unemployment data is defined to include all counts or estimates of the
total labor force, the civilian labor force, total employment, total civilian
employment, total unemployment, and total unemployment rates. The
term excludes data, obtained solely from administrative records of the
unemployment insurance system, pertaining to counts of covered
employment, the insured unemployed, and to the insured unemployment
LAUS Program Manual 1-29

rate.
b. Current Data. For the purposes of this Directive, the term current data
means the most current, complete data published by the BLS.
c. Local Area. A local area, for purposes of this Directive, is any
geopolitical unit of the United States of America and any combination or
part of any such unit or units.

LAUS Program Manual 1-30

2

Inputs to LAUS Estimation
The Current Population Survey

Introduction

T

he Current Population Survey (CPS) is a monthly survey of about 60,000 eligible
households conducted by the Bureau of the Census for the Bureau of Labor Statistics.
The survey has been conducted since the late 1940s. The CPS is the primary source
of timely information on the labor force characteristics of the U.S. population.
The CPS sample is scientifically selected to represent the civilian noninstitutional population:
that is, all persons aged 16 and over, residing in the fifty states and the District of Columbia,
not on active duty in the Armed Forces, and not living in institutions (for example, a
correctional institution or residential nursing or mental health care facility). The sample is a
State-based design, with one-fourth of the households changed each month so that no
household is interviewed for more than 4 consecutive months.
Households are interviewed for four consecutive months, are out of the sample for the next
eight, then return to the sample for the same four calendar months a year later. Respondents
are interviewed to obtain information about each member of the household 15 years of age and
older in the reference week that usually includes the 12th of the month. The responses are used
to publish data on those aged 16 and over.
The CPS provides the official employment and unemployment estimates for the nation as a
whole. In addition, it provides national data on earnings, hours of work, and other statistics.
These estimates are available by a variety of demographic characteristics including age, sex,
race, marital status, and educational attainment. Also, data are available by occupation,
industry, and class of worker. CPS Supplemental surveys collect data on a variety of topics
including school enrollment, income, previous work experience, health insurance, volunteers,
food security, and veterans.
CPS data are used by government policymakers and legislators as important indicators of our
nation's economic situation and for planning and evaluating many government programs. They
are also used by the media, students, academics, and the general public.

January 2016

LAUS Program Manual 2-1

The CPS sample is too small to provide official monthly labor force estimates for individual
states and other geographic areas. Monthly CPS employment and unemployment estimates
serve as the primary input to LAUS models which generate monthly official estimates for
census divisions, States, and selected substate areas. In addition, CPS estimates of all-other
employment, agricultural employment, and entrant unemployment are used in the estimation
for some substate areas.
Background
The Current Population Survey grew out of a program set up to provide direct measurement of
monthly unemployment, a problem that became especially pressing during the Great
Depression of the 1930s.
The Enumerative Check Census, taken as a part of the 1937 unemployment registration, was
the first attempt to estimate unemployment on a nationwide basis using probability sampling.
In addition, during the latter half of the 1930s, the research staff of the Work Projects
Administration (WPA, known prior to 1939 as the Works Progress Administration) began
developing techniques for measuring unemployment—first on a local-area basis and then
nationally. This research and experience led to the Sample Survey of Unemployment, which
the WPA began as a monthly activity in March 1940.
In August 1942, responsibility for the Sample Survey of Unemployment was transferred to the
Bureau of the Census, and its title was changed to The Monthly Report on the Labor Force. In
1948, the survey was renamed as the Current Population Survey. BLS assumed responsibility
for its publication and analysis in 1959.
The CPS is the oldest continuous household survey in the world. It has been regularly revised
and updated to keep pace with statistical and technological advances. Improvements in the
identification of households covered in the sample, sample design, methodology, and
estimation procedures have been paired with modifications to the questionnaire and interview
process to ensure increased reliability and efficiency. In 1957, the Bureau of the Census began
seasonally adjusting selected CPS data series with its X-11 model. In January 1989 the X-11
model was updated to the X-11 Auto-Regressive Integrated Moving Average (ARIMA)
method, and it was updated to the X-12 ARIMA method in January 2003. The latest version,
X-13ARIMA-SEATS, was released July of 2012.
In response to a need for more data at the subnational level, the 1970s saw a series of State
supplementary sample expansions to the CPS which, at that time, employed a national sample
design. The expansions was recognized as inefficient ways of developing State estimates. In
1985, the national-based design was changed to a State-based sampling design. This design
required that annual average State estimates fall within specified levels of reliability, while not
adversely affecting the reliability of national estimates.
Important technological advances in data collection have also been implemented. In 1994,
computer assisted telephone interviewing (CATI) and computer assisted personal interviewing
(CAPI), along with a new questionnaire design, were phased in to aid in the collection and
reliability of the data.
January 2016

LAUS Program Manual 2-2

In September 2000, the Census Bureau began augmenting the monthly CPS sample in 31 states
and the District of Columbia, as one part of the Census Bureau’s plan to meet the requirements
to produce certain estimates for the State Children’s Health Insurance Program (SCHIP).
States were identified for sample supplementation based on the standard error of their March
estimate of low-income children without health insurance. The additional 10,000 households
were added to the sample over a 3-month period. Thus, starting with July 2001 data, official
labor force estimates from the CPS and Local Area Unemployment Statistics (LAUS) program
reflected the expansion of the monthly CPS sample from about 50,000 to about 60,000 eligible
households.
Survey Process
The CPS survey process consists of three main phases: sampling, data collection, and
estimation.
Sampling involves (1) the determination, stratification, and selection of a sample of Primary
Sampling Units (PSUs) and (2) the selection of sample households within those PSUs.
Data collection involves interviewers asking households about activities during the reference
week, which is the week that contains the 12th day of the month. Labor force questions are
usually completed for each household member 15 years of age and over. BLS determines the
labor force status for each household member for the month using the information captured in
the questionnaires.
Estimation is the process of taking sample data and making estimates for the population as a
whole. Estimation involves a number of steps including data editing and imputation, basic
weighting, non-interview adjustment, ratio adjustment, compositing of estimates, and seasonal
adjustment.

January 2016

LAUS Program Manual 2-3

CPS Labor Concepts and Definitions
In the CPS, persons are classified as “employed,” “unemployed,” or “not in the labor force.”
These classifications are mutually exclusive and based on a person's labor force status during
the survey reference week (usually the week including the 12th of the month). Each person is
classified according to the activities he or she engaged in during the reference week, as defined
by the set of questions in the survey. Respondents are never asked specifically if they are
unemployed, nor are they given any opportunity to decide their own labor force status.
Interviewers do not determine a person’s labor force status. They simply ask the series of
questions and record the answers. (See Table 2-1 for the questions used to classify an
individual as employed or unemployed.)
Because the CPS is a household-based survey, it counts each person only once, at their place of
residence, even if they hold more than one job. It produces an unduplicated count of employed
and unemployed persons. In contrast, the Current Employment Statistics (CES) survey is
establishment-based and designed to produce counts of the number of jobs in the economy.
Persons holding more than one job would be counted more than once. Since the LAUS
program uses CPS concepts that reflect resident employed, adjustment must be made to make
the CES data reflect a residency basis. (See section on Dynamic Residency Ratios.)
Labor Force
The primary purpose of the CPS is to classify the civilian noninstitutional population aged 16
years and over into one of three groups: employed, unemployed, or not in the labor force.
Other information collected includes hours worked, occupation, and industry and related
aspects of the working population. It should be noted that the major labor force categories are
defined hierarchically and are mutually exclusive. Employed supersedes unemployed which
supersedes not in the labor force. For example, individuals who are classified as employed,
even if they worked less than full time, are not asked the questions about having looked for
work, and hence cannot also be classified as unemployed. Similarly, an individual who is
classified as unemployed is not asked the questions used to determine one’s primary non-labor
market activity. For instance, retired persons who are currently working are classified as
employed, even though they have retired from their previous jobs. Consequently, they are not
asked the questions about their previous employment nor can they be classified as retired.
The concepts and definitions underlying the collection and estimation of the labor force data
are presented below.
Reference week: The CPS labor force questions ask about labor market activities for one
week each month. This week is referred to as the “reference week.” The reference week is
defined as the 7-day period, Sunday through Saturday, that includes the 12th of the month. (On
occasion, the reference week in November and December may be the week including the 5th of
the month, to facilitate data collection during the holiday period.)
Civilian Noninstitutional Population (CNP): This includes all persons 16 years of age and
older residing in the 50 states and the District of Columbia who are not inmates of institutions

January 2016

LAUS Program Manual 2-4

(e.g. penal and mental facilities, and homes for the aged) and who are not on active duty in the
Armed Forces. This is the base population used in the calculation of labor force statistics.
Employed persons: This includes all persons who, during the reference week, (a) did any
work at all (at least 1 hour) as paid employees; worked in their own business, profession, or on
their own farm; or who worked 15 hours or more as unpaid workers in an enterprise operated
by a member of the family, or (b) were not working, but had jobs or businesses from which
they were temporarily absent because of vacation, illness, bad weather, childcare problems,
maternity or paternity leave, labor-management dispute, job training, or other family or
personal reasons, whether or not they were paid for the time off or were seeking other jobs.
Each employed person is counted only once, even if he or she holds more than one job.
Multiple jobholders are counted in the job at which they worked the greatest number of hours
during the reference week. (See the discussion of multiple jobholders below.)
Included in the total are employed citizens of foreign countries who are residing in the United
States and do not live on the premises of an embassy. Excluded are persons whose only
activity consisted of unpaid work around their own home (such as housework, painting,
repairing, etc.) or volunteer work for religious, charitable, or similar organizations.
The initial survey question, asked only once for each household, inquires whether anyone in the
household has a business or farm. Subsequent questions are asked for each household member
to determine whether any of them did any work for pay (or profit if there is a household
business) during the reference week. If no work for pay or profit was performed and a family
business exists, respondents are asked whether they did any unpaid work in the family business
or farm. (See Table 2-0 for the questions used to classify an individual as employed.)
Multiple jobholders: This group includes all persons who, during the reference week, had
either two or more jobs as a wage and salary worker, were self-employed and also held one or
more wage and salary jobs, or worked as an unpaid family worker and also held one or more
wage and salary jobs. A person employed only in private households (cleaner, gardener, babysitter, etc.) is not counted as a multiple jobholder, even if that person works for more than one
employer. Working for several employers is considered an inherent characteristic of private
household work. Also excluded are self-employed persons with multiple unincorporated
businesses and persons with multiple jobs as unpaid family workers.
Since 1994, CPS respondents have been asked questions each month to identify multiple
jobholders. First, all employed persons are asked “Last week, did you have more than one job
(or business, if one exists), including part-time, evening, or weekend work?” Those who
answer “yes” are then asked, “Altogether, how many jobs (or businesses) did you have?” Prior
to 1994, this information had only been available through periodic CPS supplements.
Hours of work: Beginning with the CPS redesign in January 1994, both actual and usual
hours of work have been collected. Prior to the redesign, only individuals who actually worked
less than 35 hours in the reference week were asked how many hours they usually worked. All
individuals who were at work 35 hours or more were automatically classified as full-time,
regardless of the number of hours they usually work. In the revised CPS all respondents are
January 2016

LAUS Program Manual 2-5

first asked how many hours they usually work, and are then asked in subsequent questions
about their actual hours.
Published data on hours of work relate to the actual number of hours spent “at work” during the
reference week. For example, persons who normally work 40 hours a week, but were off on
the Veterans’ Day holiday, would be reported as working 32 hours, even though they were paid
for the holiday. For persons working in more than one job, the published figures relate to the
number of hours worked on all jobs during the week.
Data on persons “at work” exclude employed persons who were absent from their jobs during
the entire reference week for reasons such as vacation, illness, or industrial dispute. Data also
are available on usual hours worked by all employed persons, including those who were absent
from their jobs during the reference week.
At work part time for economic reasons: Sometimes referred to as involuntary part-time
work, this category refers to individuals who gave an economic reason for working 1 to 34
hours during the reference week. Economic reasons include slack work or unfavorable
business conditions, inability to find full-time work, and seasonal declines in demand. Those
who usually work part time must also indicate that they want and are available to work full
time to be classified as being part time for economic reasons.
At work part time for noneconomic reasons: This group includes those persons who usually
work part time and were at work 1 to 34 hours during the reference week for a noneconomic
reason. Noneconomic reasons include illness or other medical limitation, childcare problems or
other family or personal obligations, school or training, retirement or Social Security limits on
earnings, and being in a job where full-time work is less than 35 hours. The group also
includes those who gave an economic reason for usually working 1 to 34 hours but said they do
not want to work full time or were unavailable for such work.
Usual full- or part-time status: In order to differentiate a person’s normal schedule from
his/her activity during the reference week, persons are also classified according to their usual
full- or part-time statuses. In this context, full-time workers are those who usually work 35
hours or more (at all jobs combined). This group includes some individuals who worked less
than 35 hours in the reference week—for either economic or noneconomic reasons—as well as
those who are temporarily absent from work. Similarly, part-time workers are those who
usually work less than 35 hours per week (at all jobs), regardless of the number of hours
worked in the reference week. This may include some individuals who actually worked more
than 34 hours in the reference week, as well as those who were temporarily absent from work.
The full-time labor force consists of employed persons who usually work full time and
unemployed persons who are either looking for full-time work or are on layoff from full-time
jobs. The part-time labor force consists of employed persons who usually work part time and
unemployed persons who are seeking or are on layoff from part-time jobs.
Prior to 1994, persons who worked full time during the reference week were not asked about
their usual hours. Rather, it was assumed that they usually worked full time, and they were
classified as full-time workers.

January 2016

LAUS Program Manual 2-6

Occupation, industry, and class-of-worker: For the employed, this information applies to
the job held in the reference week. A person with two or more jobs is classified according to
the job where he or she worked the greatest number of hours. An unemployed person is
classified according to their last job.
The class-of-worker classification assigns workers to one of the following categories: wage
and salary worker, self-employed worker, or unpaid family worker. Wage and salary workers
are those who receive wages, salary, commissions, tips, or pay in kind from a private employer
or from a government unit. The class-of-worker question also includes separate response
categories for “private for-profit company” and “nonprofit organization” to further classify
private wage and salary workers.
Self-employed persons are those who work for profit or fees in their own businesses,
professions, trades, or farms. Only the unincorporated self-employed are included in the selfemployed category since those whose business are incorporated technically are wage and salary
workers because they are paid employees of a corporation.
Unpaid family workers are persons working without pay for 15 hours a week or more on a farm
or in a business operated by a member of the household to whom they are related by birth or
marriage.
Occupation, industry, and class-of-worker on second job: The occupation, industry, and
class-of-worker information for individuals’ second jobs is collected in order to obtain a more
accurate measure of multiple jobholders, to obtain more detailed information about their
employment characteristics, and to provide information necessary for comparing estimates of
number of employees in the CPS and in the BLS establishment survey (the Current
Employment Statistics).
For the majority of multiple jobholders, data on occupation, industry, and class-of-worker for
the second jobs are collected only from a quarter of the sample—those in their fourth or eighth
monthly interviews. However, for those classified as “self employed unincorporated” on their
main job, class-of-worker of the second job is collected each month. This is done because
individuals who are self-employed unincorporated on both jobs are not considered multiple
jobholders.
Earnings: Information on what people earn at their main job is collected only for those who
are receiving their fourth or eighth monthly interview. This means that earnings questions are
asked of only one-fourth of survey respondents. Respondents are asked to report their usual
earnings before taxes and other deductions and to include any overtime pay, commissions, or
tips usually received. The term “usual” means as perceived by the respondent. If the
respondent asks for a definition of “usual”, however, interviewers are instructed to define the
term as more than half the weeks worked during the past 4 or 5 months. Respondents may
report earnings in the time period they prefer—for example, hourly, weekly, biweekly,
monthly, or annually. (Based on additional information collected in the interview, earnings
reported on a basis other than weekly are converted to a weekly amount in later processing.)
Data are collected for wage and salary workers, excluding the self-employed who respond that
their businesses were incorporated. Earnings data are used to construct estimates of the
January 2016

LAUS Program Manual 2-7

distribution of usual weekly earnings and median earnings. Individuals who do not report
earnings on an hourly basis are asked if they are, in fact, paid at an hourly rate and if so, what
the hourly rate is. The earnings of those who reported hourly and those who are paid at an
hourly rate are used to analyze the characteristics of hourly workers, for example, those who
are paid the minimum wage.
Unemployed persons: All persons who were not employed during the reference week but
were available for work (excluding temporary illness) and had made specific efforts to find
employment some time during the 4-week period ending with the reference week are classified
as unemployed. Individuals who were waiting to be recalled to a job from which they had been
laid off need not have been looking for work to be classified as unemployed.
A relatively minor change was incorporated into the definition of unemployment with the
implementation of the 1994 redesign. Under the former definition, persons who volunteered
that they were waiting to start a job within 30 days (a very small group numerically) were
classified as unemployed, whether or not they were actively looking for work. Under the new
definition, by contrast, people waiting to start a new job must have actively looked for a job
within the last 4 weeks in order to be counted as unemployed. Otherwise, they are classified as
not in the labor force.
As the definition indicates, there are two ways people may be classified as unemployed. They
are either looking for work (job seekers) or they have been temporarily separated from a job
(persons on layoff). Job seekers must have engaged in active job search during the last 4-week
period in order to be classified as unemployed. Active methods are defined as job search
methods that have the potential to result in a job offer without any further action on the part of
the job seeker. Examples of active search methods include going to any employer directly or to
a public or private employment agency, seeking assistance from friends or relatives, placing or
answering ads, or using some other active method. Other active methods include being on a
union or professional register, obtaining assistance from a community organization, or waiting
at a designated labor pickup point. Passive methods, which do not qualify as job search,
include reading (as opposed to answering or placing) “help wanted” ads and taking a job
training course. The response categories for active and passive methods are clearly delineated
in separately labeled columns on the interviewers’ computer screens.
Job search methods are identified by the following questions: “Have you been doing anything
to find work during the last 4 weeks?” and “What are all the things you have done to find work
during the last 4 weeks?” To ensure that respondents report all of the methods of job search
used, interviewers ask “Anything else?” after the initial or subsequent job search method is
reported.
A large number of people use the Internet to look for work. This may influence the types of
responses CPS interviewers receive to the question “What are all the things you have done to
find work during the last 4 weeks?” It is important to note that the Internet is a tool used to find
work, similar to a phone, a bulletin board, or postal mail. The use of the Internet, itself, does
not constitute a job search activity. In order to adequately categorize answers to the job search
question, CPS interviewers are trained to focus on what the respondent did on the Internet to
look for work. Many of the answers that interviewers receive about job search activities will be
January 2016

LAUS Program Manual 2-8

classified the same regardless of whether the activities were conducted over the Internet, in
person, or some other way. For example, if a respondent reports that he or she submitted an
application online, it does not matter that the activity was done on the Internet; it still should be
classified as “Sent out resumes/filled out applications.” Similarly, browsing job ads on an
Internet website is the same as looking through job ads in a printed newspaper. Both should be
classified as “Looked at ads.”
Use of the Internet presents unique challenges when coding respondents’ answers to this
question. In particular, the Internet offers a wide variety of methods for interacting with
potential employers and researching available jobs. These methods can be referred to by many
names that may change over time. Today, people email, tweet, post, update, or submit. Next
month or next year, different terms may be used. Additional instructions are provided to CPS
interviewers to address some of these challenges.
Persons on “layoff” are defined as those who have been separated from a job to which they are
waiting to be recalled (i.e., their layoff status is temporary). In order to measure layoffs
accurately, the questionnaire determines whether people reported to be on layoff did in fact
have an expectation of recall; that is, whether they had been given an indication that they would
be recalled within the next 6 months. As previously mentioned, persons on layoff need not be
actively seeking work to be classified as unemployed.
See Table 2-0 for the questions used to classify an individual as unemployed.
Reasons for unemployment: Unemployed individuals are categorized according to their
status at the time they became unemployed. The categories are:
1) Job losers: a group comprised of (a) persons on temporary layoff from a job to which
they expect to be recalled and (b) permanent job losers, whose employment ended
involuntarily and who began looking for work.
2) Job leavers: persons who quit or otherwise terminated their employment voluntarily
and began looking for work.
3) Persons who completed temporary jobs: persons who began looking for work after
their job ended.
4) Reentrants: persons who previously worked but were out of the labor force prior to
beginning their job search.
5) New entrants: persons who never worked before and who are entering the labor force
for the first time.
Each of these five categories of unemployed can be expressed as a proportion of the entire
civilian labor force or as a proportion of the total unemployed.
Prior to 1994, new entrants were defined as job seekers who had never worked at a full-time
job lasting 2 weeks or longer; reentrants were defined as job seekers who had held a full-time
job for at least 2 weeks and had then spent some time out of the labor force prior to their most
recent period of job search. These definitions have been modified to encompass any type of
job, not just a full-time job of at least 2 weeks duration.

January 2016

LAUS Program Manual 2-9

Duration of unemployment: The duration of unemployment is expressed in weeks. For
individuals who are classified as unemployed because they are looking for work, the duration
of unemployment is the length of time (through the current reference week) that they have been
looking for work. For persons on layoff, the duration of unemployment is the number of full
weeks (through the reference week) they have been on layoff.
Not in the labor force: Included in this group are all persons in the civilian noninstitutional
population who are neither employed nor unemployed. Information is collected on their desire
for and availability to take a job at the time of the CPS interview, job search activity in the prior
year, and reason for not looking in the 4-week period prior to the survey week. This group
includes discouraged workers, defined as persons not in the labor force who want and are
available for a job and who have looked for work sometime in the past 12 months (or since the
end of their last job if they held one within the past 12 months), but are not currently looking,
because they believe there are no jobs available or there are none for which they would qualify.
(Specifically, the main reason identified by discouraged workers for not recently looking for
work is one of the following: Believes no work available in line of work or area; could not find
any work; lacks necessary schooling, training, skills, or experience; employers think too young
or too old; or other types of discrimination.)
Data on a larger group of persons outside the labor force, one that includes discouraged
workers as well as persons who desire work but give other reasons for not searching (such as
childcare problems, family responsibilities, school, or transportation problems) are also
published regularly. This group is made up of persons who want a job, are available for work,
and have looked for work within the past year but not during the last 4 weeks. This group is
generally described as having some marginal attachment to the labor force.
Prior to January 1994, questions about the desire for work among those who were not in the
labor force were asked only of a quarter of the sample. Since 1994, these questions have been
asked of the full CPS sample. Consequently, since 1994, estimates of the number of
discouraged workers as well as those with a marginal attachment to the labor force are
published monthly rather than just quarterly.
Associated labor measures: Estimates of the number of employed and unemployed are used to
construct a variety of measures. They include:





Labor force: The labor force consists of the all persons 16 years of age and older
classified as employed or unemployed in accordance with the criteria described above.
Unemployment rate: The unemployment rate represents the number of unemployed as a
percentage of the labor force.
Labor force participation rate: The labor force participation rate is the proportion of
the age-eligible population that is in the labor force.
Employment-population ratio: The employment-population ratio represents the
proportion of the age-eligible population that is employed.

Table 2-0. The CPS Survey Questionnaire for Employed and Unemployed
1. Does anyone in this household have a business or a farm?
January 2016

LAUS Program Manual 2-10

2. LAST WEEK, did you do ANY work for (either) pay (or profit)?
Parenthetical filled in if there is a business or farm in the household. If 1 is “yes” and 2 is
“no”, ask 3. If 1 is “no” and 2 is “no”, ask 4.
3. LAST WEEK, did you do any unpaid work in the family business or farm?
If 2 and 3 are both “no”, ask 4.
4. LAST WEEK, (in addition to the business) did you have a job, either full or part-time?
Include any job from which you were temporarily absent.
Parenthetical filled in if there is a business or farm in the household.
If 4 is “no”, ask 5.
5. LAST WEEK, were you on layoff from a job?
If 5 is “yes”, ask 6. If 5 is “no”, ask 8.
6. Has your employer given you a date to return to work?
If “no”, ask 7.
7. Have you been given any indication that you will be recalled to work within the next 6
months?
If 1 is “no”, ask 8.
8. Have you been doing anything to find work during the last 4 weeks?
If “yes”, ask 9.
9. What are all of the things you have done to find work during the last 4 weeks?
Employed classification: Individuals are classified as employed if they answered “yes” to question 2,
or 3, or 4. For question 3, the individual must have worked 15 hours or more in the reference week or
receive profits from the business/farm).
Unemployed classification: Individuals who are available to work are classified as unemployed, if
they answered “yes” to either question 5, 6, or 7, or if they say “yes” to 8 and provide a job search
method that could have brought them into contact with a potential employer.

January 2016

LAUS Program Manual 2-11

Reliability of CPS Estimates
Because a survey is subject to nonsampling and sampling error, one cannot be a 100 percent
confident of the accuracy of the estimate, but one can estimate the likely error in a probability
sense. The larger the error the less reliable the estimate. The opposite is also true.
Nonsampling error can be attributed to many sources such as coverage, collection, and
measurement errors. A well designed and conducted survey minimizes these errors, but they
cannot be eliminated a 100 percent. (Such errors can occur with a census as well.)
Sampling error occurs because only a sample of the population has been surveyed. Standard
errors can be estimated when the probability of selection of each member of a population can
be specified. These standard errors can be used to compute confidence intervals that indicate
the range within which the true population values likely lie.
Nonsampling Error
The full extent of nonsampling error is unknown, but special studies have been conducted to
quantify some sources of nonsampling error in the CPS. The effect of the errors have been
found to be small on estimates of change, such as month-to-month change; however, estimates
of monthly levels are generally more severely affected.
Some specific types of nonsampling errors affecting the CPS include response error,
nonresponse error, error in independent population controls, processing error, and coverage
error.
Response Error: This error arises when respondent answers are incomplete (item
nonresponse) or inconsistent. It includes the inability to obtain information about all persons in
the sample, differences in the interpretation of questions, inability or unwillingness of
respondents to provide correct information, and inability to recall information. These errors are
studied by means of a reinterview program: a random sample of each interviewer’s work is
inspected through reinterview at regular intervals. This program is used to estimate various
sources of error as well as to evaluate and control the work of the interviewer. Results indicate
that the data published from the CPS are subject to moderate systematic biases.
Nonresponse Error: This error arises in situations when respondents fail to answer some or
all of the questions. In a typical month, about 10 percent of occupied sample households are
not interviewed because residents are not at home, refuse to cooperate, or are unavailable.
Therefore, sample weights are adjusted to account for households not interviewed. To the
extent that interviewed households differ from those not interviewed, the estimates are biased.
Similarly, for relatively few households, some questions are left unanswered, either because
respondents were unable or unwilling to answer or because of interviewer error. Entries for
omitted items are usually imputed on the basis of the distributions of these items for persons of
similar demographic characteristics.

January 2016

LAUS Program Manual 2-12

Independent Population Controls: These are used to account for population changes in
intercensal and post-censal years. They are extrapolated from the most recent prior decennial
Census using data on births, deaths, and net migration. Although the use of independent
population estimates in the estimation procedure substantially improves the statistical reliability
of many CPS estimates, the independent estimates are also subject to error in both the base and
the change factors. Base errors may arise in the decennial Census because of the underenumeration of certain population groups or errors in age-reporting. Also, errors in estimated
components of change since the base period of the Census affect the accuracy of intercensal
and post-censal population estimates.
Processing Error: Although the CPS employs computer-assisted interviewing and a quality
control program on coding and all other phases of data processing, some processing error is
inevitable in large surveys. Net CPS processing error is probably negligible relative to
sampling error and other nonsampling errors.
Coverage Error: Under-coverage in the CPS results from missed housing units and missed
persons within sample households. The CPS covers about 92 percent of the decennial census
population. It is known that CPS under-coverage varies with age, sex, race, and Hispanic
origin. Generally under-coverage is larger for men than for women and larger for non-whites
than for whites. Ratio-adjustment to independent age-sex-race-origin population controls,
described later in the Estimation section, partially corrects for the biases due to survey undercoverage. However, biases exist in the estimates to the extent that missed persons in missed
households or missed persons in interviewed households have different characteristics than
interviewed persons in the same age-sex-race-origin group.
Sampling Error
When a sample, rather than the entire population, is surveyed, estimates differ from the true
population values that they represent. The component of this difference that occurs because
samples differ by chance is known as sampling error, and its variability is measured by the
standard error of the estimate. When sample estimates from a given survey design are unbiased,
the sample estimate and its standard error can be used to construct confidence intervals that
contain the true population value being measured with known probabilities.
Theoretically, assuming unbiasedness, probability statements about confidence intervals are
accurate in the context of repeated sampling. Using 90% confidence as an example, if an
unbiased sampling and estimation process were repeated many times, then approximately 90%
of confidence intervals generated from that process would contain the true population value.
For a more complete explanation of confidence interval coverage refer to a standard survey
methodology text, such as Chapter 1.7 in the 3rd Edition of Sampling Techniques by William
G. Cochran (Wiley, 1977).
Examples of 90% and 95% confidence intervals are given below.
^

x = any sample estimate, such as level, rate, average, median, change over time, etc.

January 2016

LAUS Program Manual 2-13

^
^
se  x  = estimated standard error of sample estimate x
 

Then 90% and 95% confidence intervals, respectively, are calculated as
^
^ ^
 ^ 
 x  1.645* se  x  , x  1.645* SE  x  
 
 


^
^ ^
 ^ 
 x  1.96* se  x  , x  1.96* se  x  
 
 

BLS analyses are generally conducted at the 90-percent level of confidence
From the theoretical background, a practical interpretation follows: Given a 90% confidence
interval, there is an approximate 90% chance that the interval contains the true population
value. Since repeated sampling is unfeasible and the true population value is unknown, the
practical interpretation of a confidence interval is reasonable for most CPS estimates,
particularly over-the-month changes, which are less impacted by known bias and other
nonsampling errors. The practical interpretation for the 95% (or any other confidence level) is
analogous.
For estimates of change, such as over-the-month or over-the-year, confidence intervals can be
used to assess if two estimates are significantly different. Small differences relative to an
estimate’s standard error are considered insignificant, since they may have arisen due to
sampling error and not real change. If zero is included in a confidence interval of change, the
change is considered insignificant at the given confidence level. If zero is not included, the
change is considered significant at the given confidence level.
Standard error tables for national estimates are provided in the monthly publication
Employment and Earnings, with State and Regional estimates provided on the LAUS website.
CPS Monthly and Annual Reliability Criterion
Data reliability is measured by calculating the coefficient of variation (CV) of the
unemployment level; the CV is defined as the standard error of the estimate divided by the
estimate itself. The CPS sample design takes into consideration both national and State
reliability. The sample design, including the CHIP expansion, maintains a CV of 1.9 percent or
better on national monthly estimates of unemployment level. An unemployment rate of 6
percent is assumed. This means a month-to-month change in the unemployment rate must be at
least 0.2 percentage points to be considered statistically significant at a 90-percent confidence
level.
For each of the 50 states and for the District of Columbia, the design maintains a CV of 8
percent or less on the annual average estimate of unemployment level, again assuming a 6
percent unemployment rate. Due to the national reliability criterion, samples for the more
January 2016

LAUS Program Manual 2-14

populous States are substantially larger than the State design criterion requires. As a result,
annual average unemployment estimates for large States such as California, Florida, New York,
and Texas, for example, carry a CV of less than 5 percent.

January 2016

LAUS Program Manual 2-15

Sample Design
Introduction
The CPS sample design has undergone several changes throughout its history. After each
decennial census, the sample is redesigned and a new sample selected. Most changes are made
to improve the efficiency of the sample design, increase the reliability of the sample estimates,
or control cost. The current CPS sample is designed to produce reliable monthly
unemployment estimates for the nation and reliable annual average estimates for the 50 States
and the District of Columbia.
In the first stage of sampling, Primary Sampling Units (PSUs) are chosen to represent each
state. In the second stage, Ultimate Sampling Unit (USU) clusters, composed of about four
housing units each, are selected. Sample sizes and sampling rates are determined by the
predetermined reliability requirements. While the best estimates of month-to-month change
would be obtained by surveying the same households each month, indefinitely surveying a
single sample of households would inevitably lead to respondent “fatigue,” increasing the
probability of respondent refusals and errors. Therefore, a sample rotation scheme is used, with
chosen households interviewed for eight months.
Selection of Primary Sampling Units (First Stage of Sampling)
The entire area of the United States, consisting of all counties and independent cities, is divided
into 1,987 PSUs from which 852 are ultimately selected. Each PSU consists of a county or
group of contiguous counties and is defined within State boundaries.
Metropolitan areas within a State are used as a basis for forming many PSUs. Outside of
metropolitan areas, two or more counties are normally combined to form PSUs except where
the geographic area of the sample county is too large. Combining counties to form a PSU
provides greater heterogeneity; a typical PSU includes urban and rural residents of both high
and low economic levels, and encompasses, to the extent feasible, diverse occupations and
industries. Another important consideration is that the PSU be sufficiently compact so that,
with a small sample spread throughout, it can be efficiently canvassed without undue travel
cost.
Stratification and Selection of Primary Sampling Units
PSUs are separated into strata within each State. Some PSUs are self-representing and the
others are non-self-representing. Nationally, there are 506 self-representing PSUs that are
generally the most populous PSUs in each State. All the self-representing PSUs are included in
the CPS sample.
Nationally, there are 346 non-self-representing PSUs. Stratas are formed within states by
combining two or more of the remaining PSUs that are similar in such characteristics as
unemployment, proportion of housing units with three or more persons, number of persons
employed in various industries, and average monthly wages for various industries. One PSU is
selected to represent each stratum. The single PSU is called a non-self-representing PSU
January 2016

LAUS Program Manual 2-16

because it represents not only itself but all PSUs within the stratum. The probability of
selecting a particular PSU in a non-self-representing stratum is proportional to its 2010
population. For example, within a stratum, the chance that a PSU with a population of 50,000
would be selected for the sample is twice that for a PSU having a population of 25,000.
The sample of 852 PSUs will be kept for up to 10 years, the lifetime of the 2010 sample design.
Modifications to sampling procedures are made so that if needed PSUs can be restratified and
resampled before the end of the design.
Selection of Households Using Census Data
Because the sample design is State-based, the sampling ratios differ by State and depend on
State population sizes as well as national and State reliability requirements. The State sampling
ratios range roughly from 1 in every 200 households to 1 in every 3,000 households. For
example, if the state sampling ratio is 1/3,000 then that is the ratio used to select housing units
in self-representing PSUs. However for a non-self-representing PSU sampled with
probability1/2, the ratio used to select housing units in that PSU would be 1/1,500 (1/3000 = ½
x 1/1500).
Housing units for the 2010 redesign were selected from the continually updated Master
Address File. American Community Survey (ACS) block data were used to sort and stratify the
housing units in the Master Address File. Normally, census blocks are bounded by streets and
other prominent physical features such as rivers or railroad tracks. County, minor civil division
(MCD), and census place limits also serve as block boundaries. In cities, blocks can be
bounded by four streets and be quite small in land area. In rural areas, blocks can be several
square miles in size.
For purposes of sample selection, housing units are separated into two frames, a unit frame and
a group quarters frame. The unit frame mostly contains “regular” housing units with addresses
that are easy to locate (e.g., most single family homes, townhouses, condominiums, apartment
units, and mobile homes). The group quarters frame contains housing units where residents
share common facilities or receive formal or authorized care or custody.
These frames are then sampled using sampling intervals which preserve each individual State’s
sampling ratio. To reduce the variability of the survey estimates and to ensure that the withinPSU sample reflects the demographic and socioeconomic characteristics of the PSU, blocks
within the unit frame and group quarters frame are sorted using geography and block-level data
from the ACS. Examples of the variables used for sorting include proportion of minority
renter-occupied housing units, proportion of housing units with female householders, and
proportion of owner-occupied housing units.
By grouping, sorting housing units by block, and systematically sampling within blocks, the
sampling process insures that the ultimate sampling units (USUs) selected within the PSU
reflect the demographic and socio-economic characteristics of the PSU as a whole. This design
reduces the within-PSU variance, compared to the variances associated with a simple random
sample of units within the PSU.

January 2016

LAUS Program Manual 2-17

For 2010 sample design unit frame, sampling of housing units from the master address file
occurs annually for the unit frame and for the group quarters frame the sample is selected 3
years at the time. This is a departure from previous designs where a sample was selected to last
ten years that was subsequently supplemented by new construction samples. A housing unit
selected by any demographic survey conducted by census will not be available for selection by
a subsequent survey until 5 years after its last interview.
CPS State Sample Sizes and Sampling Ratios
The CPS sample of housing units is selected from within the PSUs identified above. The CPS
has a State-based sample design which allocates the sample in such a way that each of the
States and the District of Columbia has the same minimum target reliability on their annual
average estimates. A national reliability criterion is also set. Because the sample design is
State-based, the sampling ratio differs by State and depends on the various demographic
characteristics of each State. The State sampling ratios vary from approximately 1 in every 200
to 1 in every 3,000 households in each stratum of the State. In determining sample size, a
number of factors are taken into account including population density, average household size,
and variance in the unemployment rate across areas in the State.
The probability design of the CPS is self-weighting, meaning that each housing unit in a State
is given an equal chance of selection. The sampling ratio used within a sample PSU depends
on the probability of selection of the PSU and the sampling ratio for each State. In a sample
PSU with a probability of selection of 1 in 10 and a state sampling ratio of 1 in 3,000, a withinPSU sampling ratio of 1 in 300 achieves the desired overall ratio of 1 in 3,000 for the stratum.
4-8-4 Sample Rotation
The best estimates of month-to-month change would be obtained from 100-percent sample
overlap, surveying the same households every month. However, indefinitely surveying a single
sample of households would lead to respondent “fatigue” or “exhaustion” and the increasing
the probability of refusals and respondent errors.
Therefore, part of the sample is changed each month. Each monthly sample is divided into
eight representative panels, or rotation groups. A given rotation group is interviewed for a total
of 8 months, divided into two equal periods. It is in the sample for 4 consecutive months,
leaves the sample during the following 8 months, and then returns for another 4 consecutive
months. This is called a 4-8-4 panel rotation pattern. In any one month, one of the eight
rotation groups is in the first month of enumeration, another rotation group is in the second
month, and so on. Under this system, 75 percent of the sample is common from month to
month and 50 percent from year to year for the same month. (See following chart.) This
procedure provides a substantial amount of month-to-month and year-to-year overlap in the
sample, thus providing better estimates of change and reducing discontinuities in the series of
data without burdening any specific group of households with an unduly long period of inquiry.
However, the overlap creates a correlated error which must be taken into account in the Stateestimation modeling process. (See Chapter 6.)

January 2016

LAUS Program Manual 2-18

January 2016

LAUS Program Manual 2-19

Data Collection
The housing units which belong to the selected USUs (Ultimate Sampling Units) are called
“designated” households. The list of designated households is a preliminary list of potential
addresses to be sampled. Nationally, there are approximately 70,000 designated households on
this list. This list of designated household units is then refined by subtracting housing units that
have been demolished, converted to business use, relocated, or are in the sample by mistake
(i.e., units are nonresidential). The result of this refining process is a list of about 72,000
“assigned” households.
This group of assigned households undergoes further refinement when interviewers canvas the
areas removing vacant housing units, vacant sites for tents or mobile homes, units occupied by
persons with usual residence elsewhere, or units converted to temporary nonresidential use.
These are called “Type B” noninterviews. An additional noninterview type, “Type C”, occurs
when the CPS collector finds a building demolished, converted to permanent nonresidential
use, or moved from a site.
The remaining households are called “eligible” households. These are occupied by one or more
persons in scope to the CPS. There are approximately 60,000 eligible households nationally.
CPS data are collected each month during the week containing the 19th day of the month.
Respondents are asked about their labor force activity for the entire preceding week—the week
containing the 12th. A week is defined as Sunday through Saturday. The data are collected by
approximately 1,500 interviewers.
Personal visits are preferred in the first month in which the household is in the sample. In other
months, the interview generally is conducted by telephone. Approximately 70 percent of the
households in any given month are interviewed by telephone. A portion of the households (10
percent) is interviewed via computer-assisted telephone interviewing (CATI), from three
centralized telephone centers (located in Hagerstown, MD; Jeffersonville, IN; and Tucson, AZ)
by interviewers who also use a computerized questionnaire.
On the first visit, the interviewer prepares a roster of the household members and completes a
questionnaire for each person 15 years of age and older. The roster is updated with each visit.
The interviewer does not ask directly if the person is employed, unemployed, or not in the labor
force because of potential bias from the different interpretations these terms might have.
Instead, a series of questions are asked that allow a basic assignment to one of these three
categories to be made. A Computer Assisted Personal Interview (CAPI) is conducted by the
interviewers during each visit. Each interviewer has a laptop computer with a computerized
version of the CPS questionnaire. When the interviewer has completed a day’s interviews, the
data are transmitted to the Census Bureau’s central computer in Washington, D.C. Once files
are transmitted to the main computer, they are deleted from the laptops.
Of the 60,000 eligible households, about 10 percent are not interviewed in a given month due to
temporary absence (vacation, for example) of the occupants, other failures to make contact after
repeated attempts, inability of persons contacted to respond, unavailability for other reasons,

January 2016

LAUS Program Manual 2-20

and refusals to cooperate (about half of the noninterviews). Information is obtained each
month for about 55,000 housing units and 105,000 persons 16 years of age or older.
Training and Quality Control
Because of the crucial role interviewers have in the household survey, a great amount of time
and effort is spent maintaining the quality of their work. Interviewers are given intensive
training, including classroom lectures, discussion, practice, observation, home-study materials,
and on-the-job training. At least once a year, they convene for daylong training and review
sessions, and, also at least once a year, they are accompanied by a supervisor during a full day
of interviewing to determine how well they carry out their assignments.
The data collection technology and the questionnaire provide an opportunity to build functions
to assist and improve data quality into the system itself. For instance, computer-aided
interview technology guides an interviewer through complex questions: previous answers are
used to eliminate further questions that would elicit extraneous or impossible answers. Built-in
range checks for responses alert interviewers to possible inaccuracies.
Quality control procedures for the CPS are extensive, with more than 20 percent of the CPS
budget spent on training and quality control. The procedures include extensive data checking
and editing of the raw data by Census staff. Using information from the completed
questionnaire as well as additional comments provided by the interviewer in “real-time”
comments file stored in the computer, the Census Bureau staff reviews and edits the
information obtained for each person in the sample, and, where possible, identifies and corrects
omissions, unintelligible entries, and other errors.
Quality control procedures also include monitoring “on line” CATI interviews by Census
Bureau supervisory staff; a system of reinterviews, where a selection of the sample is
interviewed again, and those responses are compared with initial interview responses; and
monthly feedback to the field staff on any errors, omissions, or inconsistencies detected by the
computer edits.

January 2016

LAUS Program Manual 2-21

Weighting and Estimation Procedures
There are six main steps to the estimation process in the CPS; editing of raw data and
imputation, basic weighting, nonresponse adjustment, ratio adjustment, compositing estimates,
and seasonal adjustment. This process takes the raw data from the CPS interviews, edits it,
weights it to account for sampling probabilities, adjusts for nonresponse and consistency with
independently derived population counts for demographic sub-groups, combines current
estimates with estimates for the prior month to reduce the variability of the data, and adjusts for
seasonality.
Data Editing and Imputation
Raw CPS data are corrected for inconsistencies or missing items to make them suitable for use
in estimation. This process involves two steps: editing of the raw data and imputing for
missing or unacceptable data items.
Editing involves identifying and, where possible, correcting inconsistencies, omissions,
illegible entries, and other errors in the raw data. When the data are received at the Census
Bureau, they are reviewed for completeness and consistency. Responses to various survey
questions are interpreted and combined to classify respondents as employed, unemployed, or
not in the labor force.
Imputation involves correcting for item nonresponse—the case in which interviewed persons
do not respond to all of the survey questions or their answers to some questions are deleted
during the editing process. The empty data cells are filled using the “hot deck” method of
imputation, which is based on the premise that persons with similar characteristics provide data
that are a good approximation for the missing responses. In the “hot deck” method, data for all
interviewed persons are cross-classified by age/sex/race and geography. Missing answers are
imputed by using the data from the most recently processed record for a person in the same
age/sex/race/geography group.
Basic Weighting
The basic weighting procedure begins the process of inflating the sample data to produce an
estimate for the entire population. In basic weighting, data from each sample person are
weighted by the inverse of the probability of the person being in the sample. This is roughly
equal to the number of actual persons the sample person represents. Thus, adding the basic
weights of all sample persons having a given characteristic yields a simple estimate of the
number of persons in the population possessing that characteristic.
Nonresponse Adjustment
The weights for all interviewed households are adjusted to account for occupied sample
households for which no information was obtained because of absence, impassable roads,
refusals, or unavailability of the respondents for other reasons (Type A noninterviews). This
adjustment is made separately for clusters of similar sample areas that are usually, but not
necessarily, contained within a State. Similarity of sample areas is based on Metropolitan
January 2016

LAUS Program Manual 2-22

Statistical Area (MSA) status and size. Within each MSA cluster, there is a further breakdown
by “central city” and “balance of the MSA”. The non-MSA clusters are not split. The
proportion of sample households not interviewed is about 10 percent, depending upon weather,
vacations, etc.
Sample units found vacant, demolished, or converted to nonresidential use (Types B and C
noninterviews) are excluded from those counted for the numerator of this ratio because such
units are out of the scope of the survey. This means that the weights are not adjusted upwards.
Ratio Adjustment
The distribution of the population selected for the sample may differ somewhat, by chance,
from that of the population as a whole in such characteristics as age, race, sex, and State of
residence. Because these characteristics are closely correlated with labor force participation
and other principal measurements made from the sample, the survey estimates can be
substantially improved when weighted appropriately by the known distribution of these
population characteristics. This is accomplished through two stages of ratio adjustment, as
follows:
1) First-stage ratio adjustment: The purpose of the first-stage ratio adjustment is to reduce
the contribution to variance that results from selecting a sample of PSUs rather than
drawing sample households from every PSU in the Nation. This adjustment is made to
the CPS weights in two race cells, black and nonblack; and two age cells, 0-15 and 16+.
It is applied only to PSUs that are non-self-representing in States that have a substantial
number of black households. The first-stage ratio adjustment procedure corrects for
differences that existed in each State cell at the time of a decennial census between a)
the race distribution of the population in sample PSUs and b) the race distribution of all
PSUs. Both a) and b) above exclude self-representing PSUs. This adjustment is not
made to housing units but to the individual household-member record.
The first-stage ratio adjustment factors do not depend on response data and remain the
same from month to month during the entire intercensal period. The factors change
when a new sample of PSUs is drawn after a decennial census. The factors also change
if the non-self-representing PSU composition of a State changes for any other reason.
2) Second-stage ratio adjustment: The second-stage ratio adjustment procedure
substantially reduces the variance of the estimates and corrects, to some extent, for CPS
undercoverage at the national level. The CPS sample weights are adjusted to ensure
that sample-based estimates of population match national independent population
controls. Each month, independent estimates of various civilian noninstitutional
population distributions at the national level are produced based on the decennial census
and birth and death data from several sources. Since those characteristics are correlated
with labor force status and other items of interest, weighted CPS sample estimates are
forced to agree with the known distributions of selected population characteristics.
Beginning in 2003, the second-stage ratio adjustment includes a national and a state
coverage step, followed by three basic iterative steps. California and New York are split
January 2016

LAUS Program Manual 2-23

into substate areas (Los Angeles-Long Beach-Glendale Metropolitan Division, New
York City, and the respective balances of states). The coverage steps improve the
efficiency of adjustments for subpopulations prone to undercoverage and to account for
variations in race/gender/age differences between States.
Next, a three-step, iterative process successively is applied 10 times to adjust sample weights:
1) State step: gender-by-age cells defined for 53 States/areas.
2) Ethnicity step: 26 Hispanic and 26 non-Hispanic gender-by-age cells.
3) Race step: 34 white-alone, 26 black-alone, and 26 residual (including Asian) cells are
defined race by gender and age.
Composite Estimation
The next step in the preparation of most CPS estimates makes use of a composite estimation
procedure. The composite estimate consists of a weighted average of two estimates:
1.) The second-stage ratio estimate based on the entire sample from the current month.
2.) A composite estimate for the previous month, adjusted by an estimate of the monthto-month change based on the six rotation groups common to both months.
In addition, a bias-adjustment term is added to the weighted average to account for relative bias
associated with month-in-sample estimates. This month-in-sample bias is exhibited by
unemployment estimates for persons in their first and fifth months in the CPS being generally
higher than estimates obtained for the other months.
These composite estimates are then used as controls in the composite weighting procedure.
Both employment and unemployment are controlled in each defined cell, and not-in-labor force
(NILF) is controlled as a residual. This is an iterative process, similar to that used for secondstage weighting. Each cell in the following three steps is further split by employed,
unemployed, and not in the labor force.
1) State step: a single CPS 16+ cell is used for all 53 States/areas.
2) Ethnicity step: 10 Hispanic and 10 non-Hispanic gender-by-age cells.
3) Race step: 22 white-only, 14 black-only, and 10 Asian-only and residual gender-byage cells.
The composite estimate results in a reduction in the sampling error beyond that which is
achieved after the two stages of ratio adjustment by taking advantage of the 6/8 month-tomonth sample panel overlap of the survey. The resultant gains in reliability are greatest in
estimates of month-to-month change, although gains are also usually obtained for estimates of
level in a given month, change from year to year, and change over other intervals of time.

January 2016

LAUS Program Manual 2-24

Seasonal Adjustment
Over the course of a year, the size of the nation’s labor force and the levels of employment and
unemployment undergo sharp fluctuations due to seasonal events such as changes in weather,
reduced or expanded production, harvests, major holidays, and the opening and closing of
schools. The effect of such seasonal variation can be very large; seasonal fluctuations may
account for 95 percent or more of the month-to-month changes in unemployment.
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 nonseasonal developments, such as declines in economic activity or
increases in the participation of women in the labor force, easier to spot.
Seasonal adjustment involves using past data to approximate seasonal patterns. The seasonally
adjusted series therefore have a broader margin for error than the original data series. They are
subject to the same errors as the original series plus the uncertainties of the seasonal adjustment
process. Adjusted series are, however, useful in analyzing nonseasonal economic and social
trends.
Since 1957 CPS data have been seasonally adjusted using various versions of the Census
Bureau’s X-11 model. In January 1989 the X-11 model was updated to the X-11 AutoRegressive Integrated Moving Average (ARIMA) method. Beginning in January 2003, BLS
started using the X-12-ARIMA (Auto-Regressive Integrated Moving Average) seasonal
adjustment program to seasonally adjust national labor force data from the CPS. In January
2004, BLS converted to the use of concurrent seasonal adjustment to produce seasonally
adjusted labor force estimates. Concurrent seasonal adjustment uses all available monthly
estimates, including those for the current month, in developing seasonal factors. Previously,
seasonal factors for the CPS data had been projected twice a year. As a result of this change in
methodology, BLS no longer publishes seasonal factors for the labor force series. BLS began
using the latest version, X-13ARIMA-SEATS, in 2015.
All national labor force and unemployment rate statistics, as well as the major employment and
unemployment estimates, are computed by aggregating independently adjusted series. For
example, for each of the major labor force components—employment and unemployment—
data for four sex-age groups (men and women under and over 20 years of age) are separately
adjusted for seasonal variation and are then added to derive seasonally adjusted total figures.
The seasonally adjusted figure for the labor force is a sum of four seasonally adjusted civilian
employment components and four seasonally adjusted unemployment components. The total
for unemployment is the sum of the four unemployment components, and the unemployment
rate is derived by dividing the resultant estimate by the estimate of the labor force. Because of
the independent seasonal adjustment of various series, components will not necessarily add to
totals. (See Chapter 6 for a discussion of model-based seasonal adjustment.)

January 2016

LAUS Program Manual 2-25

CPS Data Available
Introduction
Information collected in the Current Population Survey (CPS) is made available by both the
Bureau of Labor Statistics and the Census Bureau through a broad array of publication
programs which include news releases, periodicals, and reports. This section lists some of the
different types of products currently available at BLS and Census. In some months,
supplemental questions are added to the CPS questionnaire to gather information on specific
topics. For example, August’s CPS supplement gathers data on veterans and October’s CPS
supplement collects data on school enrollment.
Bureau of Labor Statistics
The most popular publication with CPS data is the Employment
Situation news release which contains national data and is released
about 2 weeks after data collection is completed. The release includes a
narrative summary and analysis of the major employment and
unemployment developments. The news release is available at
http://www.bls.gov/news.release/empsit.toc.htm.
Another publication which is widely used is the Geographic Profile of
employment and unemployment. Annual averages labor force data from the CPS for the four
Census regions and nine Census divisions, the 50 states and the District of Columbia, selected
metropolitan areas, and selected cities are included in this publication. Data are provided on
the employed and unemployed by selected demographic and economic characteristics.
The CPS news releases, publications, and reports can be obtained at the BLS website or by
sending an email to cpsinfo@bls.gov or by calling 202-691-6378.
Table 2-1 provides a summary of the CPS data products provided by BLS.

January 2016

LAUS Program Manual 2-26

Table 2‐1. Selected BLS products from the Current Population Survey
Product 

Description 
News Releases

Periodicity 

Source 

Employment Situation 

Seasonally adjusted and unadjusted data on the Nation’s 
employed and unemployed workers by a variety of characteristics 

Monthly  

Monthly CPS 

Usual Weekly Earnings of Wage and 
Salary Workers 

Median usual weekly earnings of full‐and part‐time wage and 
salary workers by a variety of characteristics 

Quarterly

Monthly CPS; outgoing rotation 
groups 

Annual 

October CPS supplement 

College Enrollment and Work Activity  An analysis of the college enrollment and work activity of the prior 
of High School Graduates 
year’s high school graduates by a variety of characteristics 
Union Membership 

An analysis of the union affiliation and earnings of the Nation’s 
employed workers by a variety of characteristics 

Annual

Monthly CPS; outgoing rotation 
groups 

Work Experience of the Population 

An examination of the employment and unemployment experience 
of the population during the entire preceding calendar year by a 
variety of characteristics 

Annual 

March CPS supplement 

Persons with Disabilities

Labor Force characteristics of persons with disability

Annual

CPS annual averages

Families

Employment characteristics of families

Annual

CPS annual averages

Foreign Born

Labor Force characteristics of foreign‐born workers

Annual

CPS annual averages

Youth Employment

Summer employment and unemployment among youth

Annual

CPS April‐July data

Veterans

Employment Situation of Veterans

Annual

August CPS supplement 

Volunteering 

Volunteering in the United States

Annual

September CPS supplement

Displaced Workers 

An analysis of workers who lost jobs in the prior 3 years due to 
plant or business closings, position abolishment, or other reasons 
by a variety of characteristics 

Biennial 

February CPS supplement 

Tenure

Employee tenure

Biennial 

January CPS supplement

Periodicals

Employment and Earnings Online 

A monthly online periodical providing data on employment, 
unemployment, hours, and earnings for the Nation, states, and 
metropolitan areas 

Monthly 

CPS; other surveys and programs 

Monthly Labor Review 

A monthly periodical containing analytical articles on 
employment, unemployment, and other economic indicators, book 
reviews, and numerous tables of current labor statistics 

Monthly 

CPS; other surveys and programs 

Other Publications

A Profile of the Working Poor

An annual report on workers whose families are in poverty by 
work experience and various characteristics 

Annual

March CPS supplement 

Geographic Profile of Employment 
and Unemployment 

An annual publication of employment and unemployment data for 
regions, states, and metropolitan areas by a variety of 
characteristics 

Annual 

CPS annual averages

Uses of Unpublished CPS Tabulations
Unpublished tabulations such as the CPS rotation group data and the State demographic and
economic file commonly known as the DEMECON tables are provided by BLS. These
tabulations are used internally to analyze the movements in the CPS data and provide a better
understanding of the current estimates.
Rotation Groups Analysis: These data are available monthly from the CPS and represent the
raw data obtained from the monthly sample. As discussed earlier in this chapter, a given
rotation group is interviewed for a total of 8 months, divided into two equal periods. It is in the
sample for 4 consecutive months, leaves the sample during the following 8 months, and then
returns for another 4 consecutive months.
January 2016

LAUS Program Manual 2-27

Table 2-2 is an example of the monthly rotation group data.
CPS DATA FOR STATE/AREA, BY ROTATION GROUP
Civilian
noninstitutional
population

Civilian labor force
Unemployed
Total

Employed

Level

Rate

Not in
Labor
force

Sample Counts
Total 16 years and over

3,089

1,675

1,622

53

3.2

786

First month in sample

355

180

171

9

5.0

102

Second month in sample

395

231

228

3

1.3

90

Third month in sample

419

217

210

7

3.2

104

Fourth month in sample

373

191

182

9

4.7

100

Fifth month in sample

420

231

223

8

3.5

93

Sixth month in sample

386

200

193

7

3.5

106

Seventh month in sample

365

207

200

7

3.4

98

Eighth month in sample

376

218

215

3

1.4

93

Incoming rotations

775

411

394

17

4.1

195

Outgoing rotations

749

409

397

12

2.9

193

Rotations common to month before

2,314

1,264

1,228

36

2.8

591

Rotations common to month after

2,340

1,266

1,225

41

3.2

593

Rotations common to year before

1,547

856

831

25

2.9

390

Rotations common to year after

1,542

819

791

28

3.4

396

Total 16 years and over

4,304.5

2,989.4

2,887.5

101.9

3.4

1,315.1

First month in sample

501.8

330.5

311.6

18.9

5.7

171.3

Second month in sample

543.7

400.2

393.2

7.0

1.8

143.5

Third month in sample

546.6

380.0

367.3

12.6

3.3

166.6

Fourth month in sample

514.8

346.6

330.9

15.7

4.5

168.1

Fifth month in sample

575.0

415.2

399.8

15.3

3.7

159.8

Sixth month in sample

532.5

352.5

340.4

12.1

3.4

179.9

Seventh month in sample

529.1

367.1

352.8

14.3

3.9

162.0

Eighth month in sample

561.2

397.4

391.6

5.8

1.5

163.8

Incoming rotations

1,076.7

745.7

711.4

34.2

4.6

331.1

Outgoing rotations

1,075.9

744.0

722.4

21.5

2.9

332.0

Rotations common to month before

3,227.8

2,243.8

2,176.1

67.6

3.0

984.0

Rotations common to month after

3,228.6

2,245.5

2,165.1

80.3

3.6

983.1

Rotations common to year before

2,197.7

1,532.2

1,484.6

47.6

3.1

665.5

Rotations common to year after

2,106.8

1,457.2

1,402.9

54.3

3.7

649.6

Composite Weighted Counts

January 2016

LAUS Program Manual 2-28

The Composite Weighted Counts are used to compare the monthly changes in the employment
and unemployment of the sample groups. The new groups that are being introduced to the
sample are the first and fifth groups for the current month. The groups that are leaving the
sample are the fourth and eighth groups.
Table 2-3 provides an example of the spreadsheet used in rotation analysis. This spreadsheet
examines three characteristic of the monthly 4-8-4 sample rotation including how groups
entering and leaving the sample affect the monthly estimates, the labor force status of the
groups remaining in the sample, and the net effect of the entire rotation change.
The comparison of the in-coming groups for the current month to the out-going groups of the
previous months allows the analyst to determine whether a change in the employment or
unemployment level is due to an economic change in the groups currently in the sample or is
caused by the groups coming into or leaving the sample.
To calculate the in-coming versus out-going rotation change, add group 1 and group 5 of the
current month and subtract group 4 and group 8 of the previous month.
‐

‐

1

5

4

8

For example, the in-coming versus the out-going rotation change for employment in Table 2-5
is an increase of 33.8 and is derived from the sum of 311.6 (Group 1) and 399.8 (Group 5) for
the current month less the sum of 333.9 (Group 4) and 343.7 (Group 8) of the previous month.
The common rotation change is the sum of the differences between the groups present in both
the current month and the previous month. Each group of the prior month is moved up to the
next consecutive group in the current month. There are no differences for groups 1 and 5 since
these are incoming groups for the current month and did not exist in the prior month.

2
4
7

1
3
6

3
6
8

2
5
7

In the employment example in Table 2-3, the common rotation change is -12.5 and is the sum
of 11.4, -0.4, 9.4, -28.2, -4.6 and -0.1.
The net rotation change is the in-coming verses the out-going rotation change less the common
rotation change.

‐

‐

In our example the 12.5 over-the-month decrease in the employment levels of the common
rotation groups for the current month is offset by the 33.8 higher employment levels of the
incoming groups resulting in a net change increase of 21.3.
Table 2-3. CPS Rotation Group Analysis
Employment

January 2016

LAUS Program Manual 2-29

Group
1
2
3
4
5
6
7
8

Prior Month
381.8
367.7
321.5
333.9
368.6
357.4
391.7
343.7

Current
Group
Change
Month
311.6
1
393.2
11.4
2
367.3
-0.4
3
330.9
9.4
4
399.8
5
340.4
-28.2
6
352.8
-4.6
7
391.6
-0.1
8
in vs. out rotation change
33.8
common rotation change
-12.5
net rotation change
21.3

Unemployment
Group
1
2
3
4
5
6
7
8

Prior Month
4.3
9.9
19.0
15.7
11.6
12.9
5.9
17.4

Current
Group
Change
Month
18.9
1
7.0
2.7
2
12.6
2.7
3
15.7
-3.3
4
15.3
5
12.1
0.5
6
14.3
1.4
7
5.8
-0.1
8
in vs. out rotation change
1.1
common rotation change
3.9
net rotation change
5.0

Unemployment Rate
Group
1
2
3
4
5
6
7
8

Prior Month
1.1
2.6
5.6
4.5
3.1
3.5
1.5
4.8

Current
Month
5.7
1.8
3.3
4.5
3.7
3.4
3.9
1.5

Group
1
2
3
4
5
6
7
8

Change
0.7
0.7
-1.1
0.3
0.4
0.0

DEMECON Tables: These tables contain monthly and quarterly demographic and economic
data from the CPS for States and regions including the District of Columbia, New York City
and the Los Angeles - Long Beach metropolitan area.
The following tables appear in the DEMECON file:
January 2016

LAUS Program Manual 2-30

1)
2)
3)
4)
5)

Employment status of the civilian non-institutional population by sex, age, race, and
Hispanic origin
Civilians not in the labor force by sex and age
Unemployed persons by sex, age, race, Hispanic origin, and reason for unemployment
Unemployed persons by sex, age, race, Hispanic origin, and duration of unemployment
Full- and part-time status of the civilian non-institutional population by sex, age, race,
Hispanic origin

Since the data are available by age, race and sex, they are useful to gain insight on demographic
groups that may be experiencing changes contributing to the month-to-month variations in the
CPS estimates.
However, the data contained in DEMECON file are unofficial and generally do not meet the
BLS publication standards for accuracy and reliability. Tables B-2 through B-5 in the
Geographic Profiles bulletin provide generalized sampling error information for CPS annual
average data. To obtain approximate error measures for the monthly estimates in this package,
double the sampling errors in those tables.
Caution should be used in drawing inferences from these data, and they should not
be released unless they are specifically requested. When providing data to users, a
statement similar to the following should be used: The data provided are
unofficial, unpublished, data from the Bureau of Labor Statistics and do not meet
BLS publication standards for accuracy and reliability. If you publish or cite these data,
please refer to them as such.

U.S. CENSUS BUREAU
The U.S. Census Bureau reports provide information on a recurring basis
about a wide variety of social, demographic, and economic topics. In
addition, special reports on many subjects have also been produced. Most
of these reports have appeared in 1 of 3 series issued by the Census Bureau:
P-20, Population Characteristics; P-23, Special Studies; and P-60,
Consumer Income. Many of the reports are based on data collected as part
of the March demographic supplement to the CPS. However, other reports
use data from supplements collected in other months. Generally, reports are announced by
press release, and are released to the public via the Census Bureau Public Information Office.
Supplement Data Files
Public use microdata files containing supplement data are available from the Census Bureau.
These files contain the full battery of basic labor force and demographic data along with the
supplement data. A standard documentation package containing a record layout, source and
accuracy statement, and other relevant information is included with each file. The CPS
homepage is the other source for obtaining these files. The CPS homepage can be accessed at
http://www.census.gov/cps/.

January 2016

LAUS Program Manual 2-31

3

Inputs to LAUS Estimation:
The Unemployment Insurance
System

T

he Federal-State Unemployment Insurance Program provides
unemployment benefits to eligible workers who are unemployed through
no fault of their own (as determined under State law), and meet other
eligibility requirements of State law. Unemployment insurance payments
(benefits) are intended to provide temporary financial assistance to unemployed
workers who meet the requirements of State law.
Each State administers a separate unemployment insurance program within
guidelines established by Federal law. The State law under which unemployment
insurance claims are established determines eligibility for unemployment
insurance, benefit amounts and the length of time benefits. In the majority of
States, benefit funding is based solely on a tax imposed on employers.
The Federal-State Unemployment Insurance (UI) system was established in 1935
as part of the Social Security Act. It was intended by its founders to serve both as
a counter-cyclical economic stabilizer for the economy and as a central part of the
nation’s economic security system for workers with a strong attachment to the
labor force who are temporarily laid off or permanently lose their jobs. The
program is funded through mandatory payroll taxes paid by employers.
Statistics from the UI systems are the only current measure of unemployment at
the substate level available at the county and city levels. They are a key input to
the unemployment models used to estimate unemployment for the 50 States, the
District of Columbia, Los Angeles-Long Beach, the balance of California, New
York City, the balance of New York, and the five modeled substate areas and
their respective balances of state. Claims data from the UI systems are inputs to
the Handbook method for estimating labor market area (LMA) unemployment
and their use in the claims-based unemployment disaggregation yields more
accurate sub-LMA estimates than are obtained from other approaches.
While these statistics are biased for estimating total unemployment between
States in so far as they reflect the particular State's UI law, the statistics have the

LAUS Program Manual 3-1

advantage of being current and, with proper coding and tabulation, are consistent
among areas within States.

LAUS Program Manual 3-2

Federal Role
The UI system is based on a dual program of Federal and State statutes. Much of
the Federal program is implemented through the Federal Unemployment Tax Act
(FUTA). Each State administers a separate UI program with guidelines
established by Federal statute. The States also determine the eligibility, the
benefit amount and the length of time that benefits are paid.
A combination of Federal and State taxes is levied on
employers to support the UI program. The proceeds from the
unemployment taxes are deposited into the Unemployment
Trust Fund. Each State has a separate account in the Fund to
which deposits are made.
Federal UI Programs
In addition to the regular State UI program which cover the bulk of nonfarm
workers, separate Federal UI programs exist for specific types of workers.
Railroad workers receive unemployment insurance benefits through the Railroad
Retirement Board (RRB). Federal employees are covered through
Unemployment Compensation for Federal Employees (UCFE). Former military
personnel are covered through the Unemployment Compensation for ExServicemen (UCX).
Special benefits programs exist for specific situations as well. During periods of
high unemployment programs are activated that extend the time period that
individuals can receive benefits. Workers who lose their jobs as a result of the
nation's trade policies may receive special benefits after they exhaust their regular
UI benefit through the Trade Adjustment Assistance program. Workers who lose
their jobs due to a natural disaster may qualify for benefits through the Disaster
Unemployment Assistance program.
Railroad Retirement Board (RRB)
The Railroad Unemployment Insurance Act provides two
kinds of benefits for railroad employees: unemployment
benefits and sickness benefits. Benefit payments are
based on biweekly claims filed with the Railroad
Retirement Board, the Federal agency responsible for
administering the Railroad Unemployment Insurance Act.
The funds to pay unemployment and sickness benefits are provided by payroll
taxes on railroad employers only. Railroad employees do not pay unemployment
insurance taxes.
Claims filed through the Railroad Unemployment Insurance Act for
unemployment during the reference period including the 12th of the month are
used in the calculation of monthly LAUS substate estimates.

LAUS Program Manual 3-3

Unemployment Compensation for Federal Employees (UCFE)
Unemployment Compensation for Federal Employees is the benefit program for
unemployed Federal employees. Funding comes from the Federal Government
and is distributed through State agencies. Federal wages are not reported to a
State unemployment compensation agency until a claim is filed. The claimant's
Federal wages will be "assigned" to the State of the last duty station or the State
of residency if the duty station was outside the U.S or if covered work was done
in the State after leaving Federal service. This is also the case if the employer
was the Federal Emergency Management Agency (FEMA), since this is the only
Federal agency that does not report wages to the last duty station.
UCFE claims filed for unemployment in the reference period including the 12th of
the month are used in the calculation of monthly Statewide and substate LAUS
estimates.
Unemployment Compensation for Ex-Service members (UCX)
Unemployment Compensation for Ex-Service members is the
benefit program for ex-military personnel to provide weekly
income to meet basic needs while searching for employment
Those who were on active duty with a branch of the U.S.
military or active duty in reserve status as a member of a
National Guard or Reserve component continuously for 90 or
more days may be entitled to unemployment benefits based on that service. The
military wages are assigned to the State where they first file a new claim after the
separation from active duty.
Former employees of the National Oceanographic and Atmospheric
Administration (NOAA) are also covered under the UCX program. UCX benefits
are paid under the same conditions as benefits based on other employment.
However, since LAUS measures the civilian labor force, UCX claims data are not
used in the calculation of LAUS estimates.
Extended Benefits (EB) and Special Temporary Programs
Two types of UI programs grant additional weeks of unemployment
compensation to individuals who have depleted their UI benefits under the
regular UI program during economic downturns. One is the permanent program
that is funded by both the State and the Federal governments, known as State
Extended Benefits (EB), and the others are temporary programs that are financed
by the Federal government and enacted for periods of higher unemployment.
In addition, a number of States have solely State-financed programs for extending
the potential duration of benefits during periods of high unemployment, for
claimants in approved training who exhaust benefits, or for a variety of other
reasons.

LAUS Program Manual 3-4

Federal temporary extended benefits programs have been used periodically
during economic downturns since the late 1950s. The first temporary program,
Temporary Unemployment Compensation (TUC), was effective for the period of
June 1958 to June 1959. Since then, temporary extended assistance programs
have occurred seven times under different titles. These include


Temporary Extended Unemployment Compensation (TEUC), April 1961 to
June 1962;



Temporary Compensation (TC), January 1972 to March 1973;



Federal Supplemental Benefits (FSB), January 1975 to January 1978;



Federal Supplemental Compensation (FSC), September 1982 to June 1985;



Emergency Unemployment Compensation (EUC) November 1991 to April
1994;



Temporary Extended Unemployment Compensation (TEUC); March 2002 to
March 2004; and



Emergency Unemployment Compensation (EUC08), July 2008 to January
2014.
All Special Extended Benefit Programs

Name
Temporary Unemployment
Compensation (TUC)
Temporary Extended Unemployment
Compensation (TEUC)
Temporary Compensation(TC)
Federal Supplemental Benefits
(FSB)
Federal Supplemental Compensation
(FSC)
Emergency Unemployment
Compensation (EUC)
Temporary Extended Unemployment
Compensation (TEUC)
Emergency Unemployment
Compensation (EUC 08)

Effective Dates

Weeks of Benefits

06/58 – 06/59

Up to 13

04/61 – 06/62

Up to 13

01/72 –03/73
01/75 –01/78

Up to 13
Up to 13 or 26

09/82 –06/85

Up to 8, 10, 12, or 14

11/91 –04/94

Up to 7, 13, 20, 26 or 33

03/02 –03/04

Up to 13 or 26

07/08 –01/14

Up to 20, 34, 47, or 53

Source: Employment and Training Administration, Office of Workforce Security

State Extended Benefits (EB): The EB program has been a permanent part of the
Federal-State UI system since 1970 and is available to workers who have
exhausted regular unemployment insurance benefits during periods of high
unemployment as defined by the trigger “on” criteria below.
The basic EB program provides up to 13 additional weeks of benefits when a
State is experiencing high unemployment. Some States have also enacted
voluntary programs to pay up to 7 additional weeks (20 weeks maximum) of
extended benefits during periods of extremely high unemployment which is

LAUS Program Manual 3-5

defined as a State 3-month average smoothed seasonally adjusted unemployment
rate greater than 8 percent.
A State may trigger "on" for extended benefits for a week if certain criteria are
met based on the State’s insured unemployment rate (the ratio of individuals
collecting benefits to UI covered employment) or alternatively its total
unemployment rate (the LAUS smoothed seasonally adjusted rate). There are
three separate sets of criteria that can be used to determine if the EB program is
activated in a State. One is mandatory and the other two are optional and enacted
by State law.
For the mandatory trigger, a State must pay up to 13 weeks of EB if the insured
unemployment rate (IUR) for the previous 13 weeks is at least 5 percent and is
120 percent of the average of the rates for the corresponding 13-week period in
each of the 2 previous years. This comparison is called a “look back.” Ten states
have only the mandatory trigger in their State laws: Delaware, Florida, Georgia,
Iowa, Kentucky, Michigan, North Dakota, South Dakota, Utah, and Wyoming.
State law may allow for the use of one of the following alternative triggers
instead. However, if a State does not use an alternative trigger, the mandatory
method must be used to determine trigger status.
One of the optional triggers allows a State to pay up to 13 weeks of EB if the IUR
for the previous 13 weeks is at least 6 percent, regardless of the experience in the
previous years. The majority of States have specified the use of this trigger.
The other optional trigger allows a State to pay up to 13 weeks of EB if the
average total unemployment rate (TUR), smoothed seasonally adjusted, for the
most recent 3 months is at least 6.5 percent and is 110 percent of the rate for the
corresponding 3-month period in either of the 2 previous years. If such rate is at
least 8.0 percent and is 110 percent of the rate for the corresponding 3-month
period in either of the 2 previous years, the duration increases from 13 to 20
weeks. Eleven States including Alaska, Connecticut, Kansas, Minnesota, New
Hampshire, New Jersey, North Carolina, Oregon, Rhode Island, Vermont, and
Washington have permanent laws enacting the TUR options.

LAUS Program Manual 3-6

Summary of State Extended Benefits Trigger Options
Potential
Number of
Trigger Option
States*
Number
States
of Weeks
5% IUR
10
DE, FL, GA, IA, KY,
13
with look back
MI, ND, SD, UT, WY
6% IUR
without look back

31

AL, AZ, AR, CA, CO,
DC, HI, ID, IL, IN,
LA, ME, MD, MA,
MS, MO, MT, NE,
NV, NM, NY, OH,
OK, PA, PR, SC, TN,
TX, VA, WV, WI

13

TUR
(6.5% and 8%)

11

AK, CT, KS, MN, NH,
NJ, NC, OR, RI, VT,
WA

20

Total

52

* States includes DC and PR.

Source: Employment and Training Administration, Office of Workforce Security

A State triggers “off” extended benefits if, for the period consisting of the
reference week and the immediately preceding twelve weeks, the requirements of
the selected method are not satisfied.
EB and Special Temporary Programs and LAUS Estimation: Claim counts from
the EB and special temporary programs are not used in LAUS estimation. The
LAUS substate methodology already includes an estimate of unemployed
exhaustees — those persons who have exhausted the total benefit award under
State unemployment insurance within their benefit year and are still jobless. This
component is developed by using area-specific counts of individuals who
received final payments under State UI and survival rates based on the duration of
jobless spells in the Current Population Survey. (See Chapter 7.)
Claims data from permanent and temporary extensions of unemployment
compensation have not been used directly in LAUS estimation primarily because
of their temporal nature. Although individuals exhaust benefits continuously
throughout the year, insured claims under permanent or temporary extended
benefit programs for these people will be taken only when an extended UI benefit
period is established, either by trigger or by special legislation. The
administrative and methodological steps needed to incorporate such extended
benefit program data into LAUS estimation is disproportionately complex, given
the trigger on/trigger off nature of the program and the intricacies of the
temporary program. Likewise, data from the regular EB and special EB programs
are not to be incorporated into LAUS estimation. LAUS uses the estimate of
unemployed exhaustees to represent those persons who claimed their total benefit
award within the benefit year and are still jobless.

LAUS Program Manual 3-7

Trade Readjustment Allowances (TRA)
Trade Readjustment Allowances are benefits to persons whose jobs were affected
by foreign imports. Benefits are provided through the Federal Trade and the
North American Free Trade Agreement.
The Federal Trade Act provides special benefits under the Trade Adjustment
Assistance (TAA) program to those who were laid off or had hours reduced
because their employer was adversely affected by increased imports from other
countries.
The North American Free Trade Agreement (NAFTA) provides special benefits
under the NAFTA Transitional Adjustment Assistance (NAFTA-TAA) program.
Individuals who were laid off, or had hours reduced because their employer was
adversely affected by increased imports from Mexico or Canada, or because their
employer shifted production to either of these countries may qualify for benefits.
These benefits include paid training for a new job and financial help in searching
for work in other areas or relocation to an area where jobs are more plentiful.
Individuals can qualify for these special benefits only after their regular
unemployment compensation is exhausted. Thus claims data from the TAA and
the NAFTA-TAA programs are not directly counted in LAUS estimates.
Individuals who have exhausted their benefits in the regular UI program are
already accounted for through final payment counts. See UI Claims Data for
LAUS Estimation later in this chapter or Labor Market Area Unemployment in
Chapter 7 for more information on final payments.
Disaster Unemployment Assistance (DUA)
Section 410 of the Robert T. Stafford Disaster Relief and Emergency Assistance
Amendments of 1988 created a program for the payment of unemployment
assistance to individuals whose unemployment is the direct result of a major
disaster as declared by the President of the United States.
Suffering a monetary loss due to damage of property or crops does not
automatically entitle claimants to DUA. Individuals may qualify for DUA if:
they worked in or were scheduled to begin work in a county declared as a Federal
disaster county and they cannot work as a direct result of a disaster. In addition,
the work that cannot be performed must be their primary source of income and
livelihood; and they must not qualify for regular unemployment insurance from
any State.
This includes workers who suffer a loss or interruption of work as a direct result
of a major disaster, and, self-employed individuals, including farmers and day
care providers who lost or suffered a substantial reduction or interruption of selfemployment activities as a direct result of a major disaster.
Although individuals may receive DUA benefits, they are not counted as
unemployed. The definition of employed persons under the CPS includes all
those who were not working but who had jobs or businesses from which they

LAUS Program Manual 3-8

were temporarily absent because of bad weather, whether they were paid for the
time off or were seeking other jobs.
Self-Employment Assistance (SEA)
Self-Employment Assistance offers dislocated workers the opportunity for early
re-employment. The program is designed to encourage and enable unemployed
workers to create their own jobs by starting their own small businesses. Under
the program, States can pay a self-employed allowance, instead of regular
unemployment insurance benefits, to help unemployed workers while they are
establishing businesses and becoming self-employed. Participants receive weekly
allowances while they are getting their businesses started. To participate in the
program an individual must be eligible for unemployment compensation, have
been permanently laid off from his/her previous job and identified through the
State’s UI profiling system as likely to exhaust his/her benefits, and must
participate in self-employment activities including entrepreneurial training and
business counseling.
The following State have provisions for SEA: Delaware, Louisiana (law in place
but no active program), Maine, Mississippi, New Hampshire, New Jersey, New
York (expired 12/7/2015), Oregon, Pennsylvania, and Rhode Island.
Since individuals participating in a State SEA program are exempt from the State
laws relating to availability for work, search for work, and refusal to accept work,
SEA claims are not included in the claims counts for LAUS estimation.
Short-Time Compensation (STC)
The STC program, commonly known as work-sharing, provides partial UI
benefits to individuals whose work hours are reduced from full-time to part-time
on the same job. STC allows an employer, faced with potential layoffs because
of reduced workload, to reduce the number of regularly scheduled hours of work
for all employees rather than incur layoffs. Benefits are payable to workers for
the hours of work lost, as a proportion of the benefit amount for a full week of
unemployment.
The STC program currently has twenty-eight States and the District of Columbia
participating: Arizona, Arkansas, California, Colorado, District of Columbia,
Florida, Iowa, Illinois, Kansas, Maine, Maryland, Massachusetts, Minnesota,
Missouri, New Hampshire, New York, Oklahoma, Oregon, Pennsylvania, Texas,
Vermont, and Washington.
Workers are not obligated to meet the State’s regular UI requirements of
availability for work, actively seeking work, or refusal to accept work, but must
be available for their normal workweek. Thus, any claim records associated with
STC are not included in the LAUS estimation.
Office of Workforce Security Responsibilities
The Office of Workforce Security (OWS) of the Employment and Training
Administration (ETA) oversees the UI system and works closely with State

LAUS Program Manual 3-9

employment security agencies. The OWS is responsible for Program
Development and Implementation, Performance Review, Legislation, Policy and
Research, Fiscal and Actuarial services, and Information Technology.
Each Thursday, the OWS releases the “Unemployment Insurance Weekly Claims
Report” at 8:30 AM. This release can be found on the Department of Labor’s
website at www.doleta.gov.
The seasonally adjusted weekly initial claims series is a leading economic
indicator. They are the most current data on the number of people filing for
unemployment insurance. The seasonally adjusted continued claims data provide
insight as to the duration of the claims initially filed.
Interstate Statistical Data Exchange
The OWS is responsible for administrating and maintaining the Interstate
Statistical Data Exchange, which operates on the Interstate Connection Network
(ICON). The ICON is a hub-oriented data communication network that enables
54 jurisdictions, including Canada, to exchange interstate wage and benefit
transactions through batch applications. The ICON hub is located in Orlando,
Florida and maintained by Xerox State & Local Solutions.
The reporting of initial claims and continued claims (referred to as weeks claimed
by ETA) information by the Liable State to the Agent/Residence State is not only
vital for the efficient payment of interstate claims benefits, but it also has the
following important uses:
1) Interstate agent weeks claimed are needed for accurate counts of total
continued claims without earnings for LAUS estimation.
2) Interstate initial claim and weeks claimed information identifies interstate
claimants to the Agent/Residence State for purposes of providing reemployment assistance.
3) Interstate agent weeks claimed information is necessary to the
Agent/Residence State’s calculation of its insured unemployment rate that is
the trigger mechanism for the State’s Extended Benefit Program.
4) Interstate agent initial claims and weeks claimed are inputs to major economic
indicators which describe emerging and continuing unemployment conditions
in the Agent State.
The calculation of the State’s total unemployment rate (LAUS estimate), which is
the alternate trigger for the extended benefit program, includes the number of
residents that regularly commute across the State line to work in another State
and are unemployed and filing for benefits against another State. For this reason,
the reporting of commuter weeks claimed, for the survey week, is included in the
data reporting requirement of the Liable State to the Agent/Residence State.
Liable/Agent Data Transfer
The LADT record format was developed by the National Association of State
Workforce Agencies’ Interstate Benefit Committee in consultation with the
LAUS Program Manual 3-10

Unemployment Insurance (UI) Committee, the Labor Market Information
Committee and the U.S. Department of Labor’s Bureau of Labor Statistics and
the Employment and Training Administration.
The Liable/Agent Data Transfer (LADT) (Appendix/Table 3-1) is a batch
application that has a multi-purpose record format.
There are three types of records identified by the value in the Record Type field
(position 472 – Field number 62) on the LADT record layout:
1) Telephone Initial Claim (TIC);
2) Weeks Claimed (WC) (A “Weeks Claimed" record can be either an interstate
continued claim or a commuter claim);
3) Reopen/Transfer of Claim (Reopen/Transfer).
The origin of the record is identified by the value in the Liable State FIPS field
(Field number 18 - positions 184 – 185 on LADT record layout) and will be
either the alphabetic UI postal abbreviation or the numeric FIPS code.
The destination is determined by the value in the Agent State FIPS field (Field
number 20 - positions 190 – 191 on LADT record layout) and will be either the
alphabetic US postal abbreviation or the numeric FIPS code.
A commuter claim is identified in the Commuter Identification Code field (Field
number 58 - position 412 on LADT record layout). An “X” in this field indicates
a claim filed by a commuter from the Residence State, while a blank space
indicates that it is not a commuter claim.
Commuter Claim data are included in the transmittal due the first Monday of each
month. The weekending dates on commuter weeks claimed records must be
xx/12/xx through xx/18/xx only. Each month’s commuter data report must
include detail data for the “current commuter reporting month” and “prior
commuter reporting month.” “Current commuter reporting month” is the most
recently completed month. “Prior commuter reporting month” is the month
proceeding the most recently completed month.
When each Liable State’s file is received at the Hub, the records are edited and
stored. As confirmation of receipt and processing of the file, the Liable State
immediately receives, or can request output of, three reports: 1) the Liable
Summary Report; 2) the LADT Error Report; and 3) the LADT Edit Counts
Report.
After close of business on Tuesday, the Hub database is updated with each Liable
State’s file. On Wednesday morning of each week, LADT data are distributed to
the destination Agent State(s). There are two reports distributed weekly and one
monthly:
1) Agent State Summary Report – Interstate Claims (Appendix/Table 3-2) 2) Agent State Detail Data Report (contains micro data of all records)

LAUS Program Manual 3-11

3) On the first Wednesday of each month, a third report, the Agent/Residence
Commuter Weeks Claimed Report is also distributed.
(See Appendix/Table 3-3)

LAUS Program Manual 3-12

Weekly LADT Schedule
The Liable State’s detail data report is due at the Hub on
Monday of each week.

Monday
Commuter weeks claimed data must be included in the
transmission due the first Monday of the month only.
Hub processing of all data received takes place after the
close of business on Tuesday.

Tuesday

Wednesday

Hub processing of all data received takes place after the
close of business on Tuesday.
The Hub distributes Agent State detail data no later than
11:00 a.m. on Wednesday.

Additional information on the ETA, the OWS, and UI claims data can be found on the
Internet at www.doleta.gov.
Information Technology Support Center (ITSC)
In 1994, the ETA and the Maryland Department of Labor, Licensing, and Regulation
established the Information Technology Support Center (ITSC). The ITSC is a
collaboration of State employment security agencies, the Department of Labor (DOL),
and private sector partners. It promotes the appropriate application of information
technology and assists in providing States with more accurate, efficient, cost effective,
and timely service for unemployment insurance recipients.
Additional information on ITSC, UI programs, and statistics can be found on the Internet
at www.itsc.state.md.us.

LAUS Program Manual 3-13

State UI Programs
Each State law, subject to Federal requirements, establishes guidelines
determining employer coverage, individual employee eligibility, the amount and
duration of benefits paid for claims, and disqualification provisions. State UI
laws also determine the amount of payroll taxes used to fund regular UI benefits
that employers must pay. A summary of changes in individual State UI laws can
be found in each January issue of the Monthly Labor Review, published by BLS.
UI Coverage
Each State has determined its own laws regarding UI coverage, but they have
been greatly influenced by the Federal government. The Federal Unemployment
Tax Act (FUTA) provides tax incentives that have ensured States conformity with
the minimum coverage standards set down in FUTA.
In general, a covered employer is defined under the FUTA as one who has a
quarterly payroll of $1500 in the calendar year or preceding calendar year or who
employs at least one worker for at least 20 days in 20 different weeks during the
calendar year or the preceding calendar year. While many States have chosen to
expand coverage beyond the FUTA standards, the notable exceptions and
limitations are noted below.
Twenty-six weeks of regular UI benefits are provided by most States. The
following states provide an amount of benefit weeks that is not equal to 26:

State

Weeks of benefits

AR
FL*
GA*
KS*
MA†
MI
MO
MT
NC*
SC

25
12-23
14-20
16-26
30
20
20
28
12-20
20

*State has variable weeks of benefits based on its unemployment rate
based on base-period wages and unemployment rate
† Massachusetts offers 30 weeks of regular UI benefits if there is
currently no Federal emergency unemployment compensation program)

Agriculture
For the majority of States, only employers with ten or more workers in twenty
weeks, or who paid $20,000 or more in wages in any quarter, are subject to

LAUS Program Manual 3-14

unemployment insurance laws. Farm owners/operators are excluded from
coverage in all States.
Domestic Service
Private households, social clubs, and college fraternities and sororities who
employ domestic help and pay wages of $1,000 or more in a quarter are subject to
unemployment insurance laws.
Nonprofit Organizations
Coverage is required for nonprofit organizations with four or more employees in
20 weeks. Almost half of the States, however, have elected more expansive
coverage, typically covering any organization with even one employee in twenty
weeks. Ministers employed by religious organizations to perform ministerial
duties are excluded from nonprofit coverage.
Self-employed Individuals and Unpaid Family Members
As defined by the unemployment insurance laws, employment is the hiring of
workers by others for wages. Self-employed individuals are therefore excluded,
except in California, where they may elect to pay contributions for self-coverage.
Relatives are not covered unless they receive pay from the official business
payroll. However, the employment of minors by their parents, or parents by their
children, is excluded.
Railroads
Interstate railroad workers are covered by the Railroad Unemployment Insurance
Act administered by the Railroad Retirement Board. Workers on intrastate and
scenic railroads may also be covered.
State and Local Government Elected Officials and Others
All State and local government employees are covered under State UI laws with
the exception of elected officials, members of the judiciary, State national and air
national guardsmen, temporary emergency employees, and policy and advisory
positions.
Student Workers at Universities, Interns and Student Nurses
College and university students employed by the school at which they are
enrolled, such as work-study students, are excluded from coverage. Many States
also exclude the spouses of students who work at the university if the
employment is part of a program to provide financial assistance to the student.
Student nurses employed by hospitals as part of a training program are not
covered. Similarly, medical school graduates working as interns in hospitals are
excluded from coverage.
Armed Forces
Military personnel are excluded from State unemployment insurance coverage.
They are covered under a separate program, Unemployment Compensation for
Ex-Servicemen (UCX), but are not included in QCEW data. Civilian defense

LAUS Program Manual 3-15

workers, however, and all other Federal employees covered under the
Unemployment Compensation for Federal Employees (UCFE) program are part
of the data reported to the QCEW program.
Agents on Commission
Insurance and real estate agents that are paid only by commission are excluded
from coverage in almost all of the States.

LAUS Program Manual 3-16

UI Process
Just as each State has its own UI laws, each State has its own benefit payment
system for awarding UI benefits, providing documentation and fiscal control.
The benefit payment system is tied in with the taxation records system of the
State. Taxes levied from employers are deposited into the State's Unemployment
Trust Fund account.
Initial Claim
When an individual becomes unemployed, he or she must file an initial claim to
request determination for entitlement and eligibility to receive benefits.
Depending on the UI services available within each State, an unemployed person
may file an initial claim in person at an Employment Service office, over the
telephone, by mail, or via the Internet. There are three initial claim types: new,
additional and transitional.
The first claim in a benefit year filed to request benefits is
referred to as a new initial claim. A claim filed within the same
benefit year after intervening employment is call an additional
initial claim. A benefit year is the one-year period during
which an individual may receive UI benefits and is usually
related to the date of the individual's first spell of
unemployment and the filing of the claim.
A claim for benefits filed during the last week of a benefit year while a spell of
unemployment is ongoing and requesting an establishment of a new benefit year
and another eligibility determination is a transitional initial claim. A transitional
initial claim is an operational or administrative document facilitating the
transition from one benefit year to the next within a continuous spell of
unemployment. Therefore, it is excluded from the count of initial claims, since
that count represents new spells of unemployment.
Monetary Eligibility Determination
The new initial claim is evaluated, in accordance with the State’s laws, to
determine if the individual meets the monetary requirements necessary to
establish a benefit year and receive benefits and, if so, how much compensation
the individual is eligible to receive. Monetary eligibility for benefits is
determined by the amount of employment (in weeks or quarters) and wages
earned by the individual (in some combination of dollars and time worked) in a
specific base period.
A base period is a period of time prior to the benefit year (or period similar to a
benefit year) in which a claimant must have had a specified minimum amount of
insured (covered) work in order to qualify for benefits. Wages earned during this
period are used in determining the claimant’s weekly benefit amount (WBA) and
the claimant’s maximum total annual benefits. In the majority of States, the base
period is the first four quarters of the last five calendar quarters.

LAUS Program Manual 3-17

There are two types of base periods, an individual and a uniform. The individual
base period varies as to the starting date for individual claimants, while a uniform
base period starts on the same calendar year for all claimants.
A benefit year usually consists of a 1-year period or a 52-week period during
which an individual may receive annual benefits. Nearly all States have what is
called an individual benefit year in which the beginning date is dependent on the
date the claim was filed.
Each State has its own formula for computing an individual’s weekly benefit
amount (WBA), total benefit award, and duration of benefits. States may also
elect to provide benefits to dependents. See the annual ETA publication entitled
Comparison of State Unemployment Laws
(http://workforcesecurity.doleta.gov/unemploy/statelaws.asp#Statelaw) for more
details.
Claims that do not meet the State-specified requirements for the monetary
determination are denied benefits and result in a monetary disqualification.
Nonmonetary Determination
The new and additional initial claim is also subject to a nonmonetary
determination, in which a State determines whether individuals are eligible to
receive benefits based on the circumstances surrounding the loss of employment,
ability to work, availability for work, and activity in seeking work.
Each State has its own nonmonetary requirements for an unemployed individual
to receive benefits. All State laws provide that a claimant must have become
unemployed through no fault of his/her own and must be able and available to
work. The purpose of this is to provide benefits to individuals who are
unemployed primarily as a result of economic causes.
The nonmonetary determination is broken down into separation issues and
nonseparation issues.
Separation issues refer to situations surrounding the termination of the
employment relationship. This includes incidences where the individual
voluntarily quits without good cause, or voluntarily quits for personal reasons.
(For LAUS purposes, these individuals would be counted as unemployed even
though they do not qualify for UI benefits, as long as they are willing and able to
work, and are actively seeking employment.)
Nonseparation issues pertain to situations in which the individual's actions, the
type or seasonality of occupation, or income preclude eligibility. An individual
can be precluded from receiving benefits if they are not willing or able to seek
employment, or if they refuse suitable employment. Seasonal employees, such as
school personnel and professional athletes, are not eligible for benefits during the
time period between terms of employment. Disqualifying incomes includes
pensions, severance pay and other UI compensation, such as EB, TRA or DUA
benefits. Individuals disqualified because they were not able or willing to work
would not be counted as unemployed for LAUS purposes. Those disqualified for

LAUS Program Manual 3-18

receiving UI because they have income from pensions or severance pay would be
counted as unemployed, as long as they are willing and able to work, and are
actively seeking employment and the amount of money earned does not exceed the
weekly benefit amount.
An individual may pass the monetary eligibility requirements but may not receive
benefits in the event of voluntary leaving without good cause, discharge or
suspension or misconduct, refusal of suitable work, labor dispute, and false
statements.
A claimant who does not meet the nonmonetary requirements and is denied
benefits results in a nonmonetary disqualification.
Nonmonetary Penalties
The circumstances of a nonmonetary disqualification may preclude the individual
from ultimately receiving benefits.
Separation Issues:
Voluntarily Leaving Work: Since the UI program is designed to compensate wage
loss due to lack of work, voluntarily leaving work without good cause is an
obvious reason for disqualification from benefits. All States have such
provisions. In most States, the disqualification lasts until the worker is again
employed and earns a specified amount of wages. However, in a few States the
disqualification is a fixed number of weeks and can be up to 12 weeks depending
on the reason why the individual needed to leave work. In addition, some States
may also reduce the individual’s benefit rights, usually equal in extent to the
weeks of benefit postponement.
Discharge for Misconduct: A number of States have a variable disqualification
for discharge for misconduct. In some States the range is small, for example, the
week of occurrence plus 3 to 7 weeks. In others, the range is large, 5 to 26
weeks. Several States provide a fixed disqualification, and others disqualify for
the duration of the unemployment, or longer. Some States may reduce or cancel
all of the worker’s benefit rights and some provide for disqualification for
disciplinary suspensions.
Non-separation Issue:
Refusal of Suitable Work: Several States disqualify for a specified number of
weeks (3 to 20) any workers who refuse suitable work; others postpone benefits
for a variable number of weeks, with the maximum ranging from 1 to 12.
Monetary Penalties
The penalty for a monetary disqualification is in effect until the individual
becomes subsequently employed and earns a specified amount of wages to
become eligible.
Appeals
An individual whose claim has been denied for either
monetary or nonmonetary reasons may request an
opportunity for a fair hearing before a UI tribunal or UI

LAUS Program Manual 3-19

authority. A request can be made for a review by an appeals authority on the
State's determination of the claim for benefits, the employer's contribution rate, or
a decision made by a lower appeals authority.
The employer for whom the individual worked during the wage qualifying base
period is charged with the liability for the claim payment. This employer may
also challenge the UI decision of the individual's eligibility to receive benefits.
Continued Claim
Once a claimant passes the monetary (for new initial claims only) and
nonmonetary eligibility requirements, the individual must satisfy mandatory
requirements for each week of unemployment for which he/she claim benefits.
These weekly requirements include actively seeking employment and being
available for work. Certification, or certifying, is the form and process by which
an individual attests to the facts that determine eligibility for a given week.
This certification process must be completed for each week that the individual
claims benefits. Most States establish a 52-week period (benefit year) during
which the individual may submit claims for benefits.
Individuals may also receive earnings from regular employment, or odd jobs,
while certifying for a week of unemployment. In such cases, these claims are
designated as continued claims with earnings. They are not used in LAUS
estimation since they do not meet the LAUS definition of unemployed. (Even
one hour of work results in the classification of the individual as “employed” in
the CPS.) Continued claimant counts without earnings due to employment are
the primary source of unemployment inputs for the Handbook method and an
important input into State models.
Final Payments
A final payment is the last continued claim for which an individual can receive
benefits within a benefit year. At this point, the individual has exhausted the
maximum benefits as was calculated at the time of filing and monetary
determination.
Final payment recipients, also referred to as UI exhaustees, are the primary input
to the monthly estimate of unemployed exhaustees of the Handbook method. (See
Chapter 7)
Benefits delivery arrangements
The State conducts UI activities under all of the following arrangements:
Intrastate Benefit Arrangements.
The State provides benefits to individuals who reside in and worked in that State.
The State also provides benefits to individuals who worked (and would continue
to seek work) in the State but reside in a border State. Intrastate claims filed in
the State where the claimant worked but does not reside are referred to as
commuter claims.

LAUS Program Manual 3-20

Interstate Benefit Arrangements.
To encourage a claimant to move from a State where no suitable work is available
to one where there is a demand for the type of service that the claimant is able to
render, States have made agreements to protect the rights of workers who make
such moves. These arrangements permit the collection of benefits from the State
in which an individual has qualifying wages (liable State) even though the
claimant is not physically present in that State. The State in which the individual
is located may accept the claim, acting as Agent for the State that is liable for the
benefits claimed. The liable State may also accept the claim directly from the
claimant by telephone or internet means. Determinations on eligibility,
disqualifications, and the amount and duration of benefits are made by the liable
State.
Wage-Combining Arrangements.
This arrangement permits workers to combine their wages and employment in
more than one State and file in a single State. This holds for situations where
there are insufficient wages and employment to qualify for benefits in any one
State and where, having sufficient wages and employment to qualify for benefits
in one State, the benefit amount would be increased by combining wages and
employment in other States. In addition, this arrangement permits workers
having sufficient wages and employment to qualify for benefits in more than one
State, to combine their wages in those and any other States in which they had
wages and employment in the base period of the liable State.

LAUS Program Manual 3-21

UI Claims Process
The following chart illustrates the process of claims validation in a State.

LAUS Program Manual 3-22

UI Claims Data for LAUS Estimation
Counts of individuals associated with continued claims are used in the
development of State and model-based area unemployment. Counts of
individuals associated with two types of claims, continued claims and final
payments, are used in the development of sub-state unemployment estimates.
Continued Claims
Continued claimant data for certifications to unemployment in the week including
the 12th of the month from the State UI program are used each month in the
development of State and area model-based total unemployment. The State UI
program data, along with data from the UCFE, and RRB programs, are used each
month to calculate monthly LAUS estimates for sub-state areas. Because the
unemployment measurement is limited to the labor force status of the civilian
population, claims data from the UCX program are excluded.
The continued claimant count is made up of persons who certified to a
compensated or non-compensated week of unemployment.


Compensated claims relate to those claimants who are receiving UI
benefits. A claimant who has worked during the week and received any
earnings while certifying for unemployment does not meet the criteria for
the CPS definition of unemployed and is omitted from the count used in
LAUS estimation even if the earnings do not result in a reduction in the
benefit award.



Noncompensated claims include individuals who are unemployed and are
certifying weekly for benefits but are not receiving compensation for any
of the following reasons:
o They are certifying during the waiting week,
o They are certifying while appealing a monetary or nonmonetary
disqualification. Or
o They are certifying while filing a pending claim.

These individuals are unemployed, are in the UI system, and are included in the
continued claimant count for LAUS estimation.
The LAUS program recognizes three types of continued claims data: intrastate,
commuters and interstate for both regular State UI and UCFE programs.
Intrastate claims
Intrastate claims are claims filed by unemployed persons in the State where they
live and worked.
Commuter claims
Commuter claims refers to claimants who worked and would continue to seek
work in one State while living in close proximity in another State. These

LAUS Program Manual 3-23

claimants are treated as if they reside in the State of prior employment and file
intrastate claims in the State in which they had worked.
Interstate
Interstate refers to interstate claims filed by claimants who resided and became
unemployed in one State, and, during their spell of unemployment, moved to
another State and filed for UI benefits in the new State of residence. The State
where the claimant first became unemployed is still liable for that spell of
unemployment and the UI benefits that the claimant receives.
Interstate and commuter claims data are exchanged through the ICON system,
which enables States to transmit, and to retrieve appropriate claims data. The file
format used by ICON is called Liable-Agent Data Transfer (LADT).
Final Payments
Final payment recipients, also called UI benefit exhaustees, are continued
claimants who file for a week of unemployment that exhausts their total benefit
award.
Once claimants receive their final payment and leave the UI system, States are
unable to track them. The LAUS methodology estimates the number of UI
exhaustees who are still unemployed by applying a survival rate to the number of
final payments. This survival rate is developed from CPS data on duration of
unemployment. For further details regarding this process, see the Labor Market
Area Unemployment section of Chapter 7.

LAUS Program Manual 3-24

Differences: UI Data versus the CPS
CPS data are used directly to produce official labor force estimates for the nation
and they are the key input to LAUS model based estimates of unemployment for
States and selected areas. Differences between the State UI count of continued
claimants without earnings and CPS unemployed result primarily from
differences in program coverage of the unemployed by the UI system.
Certain industries and occupations are excluded from UI coverage, including:


employees of certain nonprofit organizations;



insurance and real estate agents on commission;



agricultural workers on small farms and certain seasonal farm workers;



some domestic workers;



self-employed persons;



unpaid family workers;



some State and local government employees;



student nurses and interns in hospitals; and



railroad workers (covered under the RRB program).

Certain unemployed individuals may not be able to receive UI benefits,
including:


people who have not worked long enough and therefore have insufficient
wages to establish eligibility for benefits under UI;



people who quit a job or were discharged for misconduct;



people who have exhausted their benefits and could not re-establish a
benefit year; and



people with no recent earnings, such as new entrants or reentrants to the
labor market.

The UI exclusions limit program coverage to unemployed individuals with recent
employment experience and exclude unemployed new entrants and reentrants to
the labor market. Also, much of agricultural unemployment is not represented in
UI statistics.

LAUS Program Manual 3-25

BLS Standards for UI Data
LAUS labor force estimates produced for State and substate levels use the same
definitions and concepts as the CPS so that the resultant estimates are consistent
with the CPS and comparable within and across States. In order to further ensure
this, BLS sets standards for the UI statistics used in LAUS estimation
Standards for Continued Claims and Final Payments
Two insured unemployed counts, continued claimants and final payment
recipients, are used in the development of LAUS substate unemployment
estimates. For the model-based unemployment estimation procedure, only
continued claimants are used.
Continued claimants are persons certifying to a compensated or noncompensated
week of unemployment under the State UI and UCFE programs. Because
measurement is limited to the labor force status of the civilian population, the
UCX program is excluded. The continued claimant count includes intrastate
claimants, commuter claimants (based on State of residence), and interstate
claimants (based on State of residence). UCFE program data are not used in
developing unemployment estimates for modeled areas.
Persons receiving final payments are continued claimants certifying to a week of
unemployment which represents the last regular benefit payment in the benefit
year. Further benefits are not available until the beginning of a new benefit year.
The BLS standard of quality for these continued claims and final payment counts
is as follows:


the counts reflect the State and county of residence of the unemployed;



the counts are unduplicated and based on the Social Security number and the
claimant’s week of unemployment for which the claimant certified;



the counts include both compensated and noncompensated claimants as
described above;



for continued claimants, the claimant’s week of certification is consistent with
the CPS reference week., i.e., the week including the 12th of the month(see
also the December reference week below);



for persons receiving final payments, the counts are weekly, based on the
week for which the claimant is certifying;



the counts exclude persons with any earnings due to employment, regardless
of their entitlement to full weekly UI benefits.

LAUS Program Manual 3-26

Reference Week for December and November
Normally, the reference period is the week including the 12th of the month.
However, this is may not always be the case for the months of December and
November.
In December, the actual reference week used by the CPS and, thus, LAUS,
depends on the number of business days between the 12th of the month and
Christmas day. As a result, the reference week is often the week that includes the
5th of the month, the week prior to the week that includes the 12th.
Similarly for November, if Thanksgiving occurs in the week of the 19th, then the
reference week will be the week including the 5th.
Moving the reference period up a week allows adequate time for CPS data
collection and processing prior to the Christmas and Thanksgiving holidays. The
change in the reference week is also necessary because CPS response rates fall
substantially during the days immediately before these holidays.
Standards for Initial Claims
While not part of direct LAUS estimation, initial claims are integral to the UI
process itself. Also, initial claims counts may be used in atypical or exception
procedures in the Handbook method to develop an estimate of those unemployed
who are eligible for UI but delay filing or never file for unemployment benefits.
(Estimates of delayed and never filers are not used in Handbook estimation, but it
is useful to define the standards for initial claims.)
An initial claim is a notice filed by an individual to request determination of
entitlement to and eligibility for compensation. A new initial claim is the first
claim filed by the claimant within the benefit year. An additional initial claim is a
second or subsequent claim filed by the claimant within the benefit year after an
intervening period of employment. .
Reference Period
Unlike continued claims that relate to a certification period in the past, initial
claims do not refer to a reference period, but rather represent a point in time.
Information requested on the initial claim form typically includes the date of
filing, and the date of separation and the separating employer.
Excluded Groups
For purposes of defining spells of insured unemployment, the following types of
initial claims are excluded from consideration:
1) Invalid new initial claim where the individual is found to be monetarily
ineligible for UI.
2) Transitional initial claim, where a new, unique spell of unemployment has not
occurred. Such an initial claim is filed by a claimant during a spell of
LAUS Program Manual 3-27

unemployment in the last week of his/her benefit year, requesting an eligibility
determination and establishment of a new benefit year. Because the claimant is in
a continuous spell of unemployment and is also filing a continued claim, such
transitional initial claims are excluded from the count representing new, emerging
unemployment.
3) Reopened claim, where a claimant reopens a continued spell of unemployment
after ceasing certifying to unemployment and withdrawing from the labor force.
If this atypical action does not reflect an intervening spell of employment, the
State may administratively reopen the claims series and allow the claimant to
resume filing continued claims. These claims are not to be considered initial
claims.
Standards for Residency Coding
With Federal funds allocated to areas below the State level, the use of claims data
by residence is imperative, not only as a determinant of the labor market area’s
total unemployment estimate, but also in the development of county and subcounty estimates. These estimates are created through a method called
disaggregation. For further details regarding this process see Chapter 9.
The residency requirement for claims data calls for the coding and tabulating of
claimants by State, county of residence and place of residence. The geographic
requirement applies to areas within the State paying the benefits (or acting as
agent State for interstate claims) and to counties in contiguous States whose
residents cross State lines to file intrastate claims in the State holding their wages
and paying the benefits (commuter claimants).
The UI Database Survey
Beginning in 1975 an effort was undertaken by BLS to improve the UI data used
in LAUS estimation. This effort started with a survey of the State UI database
systems. There were two primary reasons for the survey. The first was to
describe the methods States were using to obtain the UI statistics used in
developing LAUS estimates, and the second was to provide input to determining
the necessary modifications to State systems to achieve more uniform UI data
series. Based on this survey, the BLS standards for UI statistics were developed.
A plan of action was developed to eliminate inconsistencies in UI statistics both
between and within States as compared to official labor force concepts. Claimant
data that represented an unduplicated count of individuals by State and county of
residence for the appropriate reference period and with maximum adherence to
the CPS definition of unemployment in terms of any earnings due to employment
in the week of certification was the focus. These characteristics—unduplicated
count of persons, residency, reference period, and exclusion of persons with any
earnings—are essential elements for the UI claimant statistics used in
unemployment estimation, and areas where improvement efforts have been
concentrated.

LAUS Program Manual 3-28

To obtain an unduplicated count of persons, Social Security numbers are used.
Also, the use of the week of certification prevents multiple counting because an
individual can certify only once to a week of unemployment.
Basing the UI statistics on claimant residency rather than more program-related
locations such as local office or place of employment ensures correspondence
with official labor force concepts. Proper residency coding affects all LAUS uses
of UI data.
The requirement that UI statistics relate to the week including the 12th of the
month also ensures correspondence with the official labor force estimates. The
exclusion of persons with any earnings makes the data consistent with the CPS,
where one hour of pay qualifies an individual as employed.
The initial UI Data Base Survey was a three-day-long meeting of State research,
State UI, and BLS staff. A detailed description of UI administrative and
operational processes was obtained.
Based on the results of the UI Data Base Survey, contracts were established with
all States to develop unique monthly statistical counts that reflect the BLS
standards for UI statistics and that are used for LAUS estimation purposes. These
State-developed tabulations have been replaced by tabulations developed by the
Program to Measure Insured Unemployed Statistics (PROMIS).
The UI Claims Review and Validation Project
In 2001 a simplified form of the UI Database Survey was introduced. Entitled the
UI Review and Validation Project, this examination of the data sources for UI
inputs into LAUS was designed to be less of a burden for the States to participate
in and to be conducted more frequently than the UI Database Survey. Due to the
ever-changing nature of technology, UI laws, and administrative procedures, it is
necessary to re-examine the data sources and procedures on a regular basis.
The project included reviewing and validating the quality of claimant data
extracted from the unemployment insurance (UI) system databases in the States.
In addition to the LAUS program, the Mass Layoff Statistics (MLS) program
required UI statistics on initial claims for the event trigger, continued claims for
describing the ongoing layoff, and final payments. Thus, the validation of UI
data extract parameters and specifications used in the State were integral to
quality assurance of not only the UI claims data inputs but also the LAUS
estimates and the MLS statistics.
The validation procedure consisted of two separate stages, an initial examination
of the current procedures in place for obtaining UI claims data for use in LAUS
and MLS estimations and a follow up session to address any problems identified
in the initial examination.
The initial examination involved an on-site visit to the State agency by BLS staff.
The purpose of the visit was to interview the State staff who were responsible for
processing and extracting UI claims data for LAUS and MLS estimation and to
conduct a detailed examination of the UI source files and the extract programs

LAUS Program Manual 3-29

used for LAUS and MLS purposes. A standard questionnaire form was used to
document the claimstaking method, the filenames where the claims are stored,
and to determine if the extract program was capturing all the necessary claims
data input.
Upon completion of the questionnaire, examination of the UI source files, and a
review of the extract program, a determination was made if the State's claims data
inputs for LAUS and MLS meet the BLS standards for UI claims data. If any
shortcomings were discovered, then a second stage was initiated to ensure that
any deficiencies will be corrected.
In the second stage, problems identified in the initial inquiry were addressed. The
BLS and the State staff determined solutions to ensure that the BLS data
standards are met. Solutions included the re-writing of UI claims extract
programs or other programs that process claims data for LAUS and MLS. In
addition, the BLS and State staff determined the necessary resources needed to
implement the corrections.
The validation process was an on-going project with each State being certified
and re-validated periodically as UI polices and technologies change. Two events
led to the elimination of this project. One was the limitation of travel to States
due to budget restraints starting in 2008. The other was the increasing number of
States implementing the PROMIS system.
The in-depth research and analysis required by States to gain approval to
officially use PROMIS has provided a much more comprehensive examination of
the participating State’s UI system and claims extract programs. In many cases
the set up of new UI database extract programs specifically for PROMIS has
revealed inaccuracies hidden in the legacy system’s extract program that went
unknown and were not uncovered by the validation project. The PROMIS
approval process has effectively removed the need for the UI Review and
Validation Project at the present time.
Program to Measure Insured Unemployment Statistics (PROMIS)
The PROMIS system is a stand-alone PC-based system that stores
all claimant information, including socioeconomic characteristics,
and generates the UI inputs for LAUS estimation. In addition to
generating input files, PROMIS can be used to develop tabulations
at the State and area level of UI claimants by socioeconomic
characteristics. PROMIS operates as the clearinghouse for multi-purpose input
data, allowing flexibility to provide a more complete picture of the
unemployment situation at substate levels.
The PROMIS system is designed to provide States with increased quality
assurance and resource efficiency to develop monthly statistics. The
implementation of PROMIS by a State involves the creation of a new UI database
extract program. This enables the most up-to-date LAUS criteria to be
incorporated into the extract program and for the program to be written in a
current programming language. The extracts used by most States for their legacy
LAUS Program Manual 3-30

systems were developed decades ago and were often not updated as needed.
These legacy extracts were written in programming languages that are now
antiquated and thus are confusing and unfamiliar to current State staff. A
rigorous examination of the PROMIS data and the legacy system data, which is
also required, has often exposed flaws in the legacy extract that went unnoticed.
The PROMIS data quality is further enhanced by the Residency Assignment
System, which corrects erroneous address information and assigns geocodes for
States, counties, cities and towns. (See the Residency Assignment System
section of this chapter on pages 3-29 to 3-47 for more information.) In addition,
the use of PROMIS facilitates the implementation of claims-based disaggregation
by providing claims data at the city and town level. (See Chapter 9 on
Disaggregation for more details.)
The following files are produced by the PROMIS system:











City Claims Input Files compiled by Counties and Cities
State UI and UCFE Continued Claims Less Earnings (batch IDs: M03,
M04, M05, M06, M07, M08) for the LAUS State System
State UI and UCFE Continued Claims Less Earnings (batch IDs: M03,
M04, M05, M06, M07, M08) compiled by Labor Market Area (LMA) for
New England States
Week 1 through Week 6 Final Payments (batch IDs: M10, M11, M12) for
the LAUS State System
Week 1 through Week 6 Final Payments (batch IDs: M10, M11, M12)
compiled by Labor Market Area for New England States
All UI and UCFE Continued Claims Less Earnings (batch IDs: M03,
M04, M05, M06, M07, M08), and Week 1 through Week 6 Final
Payments (batchM03, M04, M06, M07, M10, and M11
All UI and UCFE Continued Claims Less Earnings (batch IDs: M03,
M04, M05, M06, M07, M08), and Week 1 through Week 6 Final
Payments (batch IDs: M10, M11, M12) compiled by Labor Market Area
(LMA) for New England States, and one large file containing the output
for batch IDs M03, M04, M06, M07, M10, and M11
MicroData detail behind all LAUS output data (batch IDs: M03, M04,
M05, M06, M07, M08, M10, M11, M12, C06 if New England State, C07
if non-New England State)
12-month LSS output files for LAUS annual processing: one consolidated
file containing non-commuter output (batch IDs: M03, M04, M06, M07,
M10, M11); one file containing City Claims (batch IDs: C06 if New
England State or C07 if non-New England State); one consolidated file
with each of the commuter outputs (batch IDs: M05, M08, M12)

For more information, see the PROMIS system user’s guide available on the
LAUS/MLS Intranet under Operations and Manuals.

LAUS Program Manual 3-31

Residency Assignment System

To assist States in correctly coding the residence of their UI claims
records, the LAUS National Office has made available address
correction and geocoding software called the Residency Assignment
System (RAS). RAS corrects erroneous addresses, city, state and ZIP
codes and assigns geocodes for States, counties, incorporated places, and minor
civil divisions (MCD). It can also be customized to assign State specific codes
and other geocodes available from the US Census Bureau.
States provide the input file(s) and corresponding format file(s) that defines the
layout of the input file(s). States determine the layout of the output file and
which geocodes are to be assigned. Geocodes include State and county FIPS
codes, FIPS place codes, minor civil divisions (MCD) codes, longitude and
latitude, and census tracts and blocks. In addition, other codes can be added such
as State specific codes or other census geocodes. Project files are created by
national office staff to reflect the output file layout and the geocoding
requirements of each State.
States upload their input file via EUSweb to their State folder in either the LAUS
directory or the PROMIS directory (if the State is approved for PROMIS). Each
State can upload, download, rename or delete any of their files in their State
folder. The system will automatically process the input file as long as the file is
appropriately named (see the following section on input files).
A polling agent program searches each State folder for a new input file and
automatically runs RAS when one is detected. Once the input file is processed,
RAS places the output files in the State folder and saves the input in the Archive
subfolder located within the State folder. The polling agent also notifies the State
via email when the file has been processed. If for any reason the same file has to
be processed again, the original input file can be moved or copied from the
archive folder back into the State folder, and the system will automatically
process it. The RAS output includes the processed file (STout.txt) with the
correct addresses and geocode assignment.

LAUS Program Manual 3-32

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout
FLD
NBR

1

2

FIELD NAME

Social
Security
Number
Claimant’s
Name - First

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

W/C IB

W/C COMM

N

1

9

Y

Y

Y

Y

A/N

10

12

Y

Y

DESCRIPTION

Enter Claimant’s
Social Security
Number
Enter the claimant’s
first name. First
position cannot be
blank. Enter at least
one alphabetic
character.
Claimant’s middle
initial.

3

Claimant’s
Name Middle Initial

A/N

22

1

Y

Y

4

Claimant’s
Name - Last

A/N

23

23

Y

Y

5

Mailing
Address Street
Mailing
Address - City
Mailing
Address State

A/N

46

30

Y

Y

Y

Y

Enter Claimant’s (Mailing) Street

A/N

76

19

Y

Y

Y

Y

A/N

95

2

Y

Y

Y

Y

Enter Claimant’s (Mailing) City
Enter Claimant’s (Mailing) State

Mailing
Address - Zip
Code
Residence
Addr - Street

A/N

97

9

Y

Y

Y

Y

Enter Claimant’s (Mailing) Zip Code

A/N

106

30

6

6

6

6

Enter Claimant’s (Residence) Street

Residence
Addr - City

A/N

136

19

6

6

6

6

Enter Claimant’s (Residence) City

6
7

8

9

10

Enter the claimant’s
last name. First
position cannot be
blank. Enter at least
one alphabetic
character.

LAUS Program Manual 3-33

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

W/C IB

W/C COMM

DESCRIPTION

11

Residence
Addr State

A/N

155

2

6

6

6

6

Enter Claimant’s (Residence) State

12

Residence
Addr - Zip
Code

A/N

157

9

6

6

6

6

Enter Claimant’s (Residence) Zip
Code

13

Claimant’s
Telephone
Number

N

166

10

Y

Y

14

Year of
Birth

N

176

4

Y

Y

Y

Y

Claimant’s Year of
Birth - Format is
“CCYY”.

15

Sex

N

180

1

Y

Y

Y

Y

Enter the sex of the
claimant
1 = Male
2 = Female
3 = Unknown

16

Race

N

181

1

Y

Y

Enter Area Code,
Exchange, and
Extension of the
Claimant’s
Telephone Number

Claimant’s Race
Code
1 = White
2 = Black
3 = Asian
4 = American
Indian/Alaskan
Native
5 = Native
Hawaiian/Other
Pacific Islander
6 = Information Not
Available

LAUS Program Manual 3-34

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

17

Education

N

182

2

18

Liable
State FIPS

A/N

184

2

Y

Y

19

Liable
State
Office/Call
Center
Number

N

186

4

Y

Y

W/C IB

W/C COMM

DESCRIPTION

Highest Grade
Completed
01 - 12 Actual grade
completed (12 = GED)
13 = 1 year of college or
technical school
14 = 2 years of college
or Associate
degree/technical school
15 = 3 years of college
16 = 4 years of college
or undergraduate degree
17 = 1 year of post
graduate study
18 = 2 years of post
graduate study or
Masters degree
19 = Doctorate

Y

Y

Liable State FIPS Code.
The Liable State cannot
be the same as the
Agent State.

Enter number that
identifies the Liable
Interstate office/Liable
Call Center that handles
the claim.

LAUS Program Manual 3-35

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

W/C IB

W/C COMM

DESCRIPTION

21

Agent State
Local
Office/Call
Center
Number

N

192

4

Y

Y

Y

Y

Enter number that
identifies the Local
office/Call Center where
the claimant filed the
claim.

22

Residence
State FIPS

N

196

2

5

5

1

5

Residence State FIPS
Code. The Residence
State cannot be the
same as the Liable
State.

23

Residence
County
FIPS

N

198

3

Y

Y

2

2

Residence County FIPS
Code.

24

Residence
City/Town
FIPS

N

201

4

Y

Y

Y

Y

Residence City/Town
FIPS Code.

25

Date Claim
Taken

N

205

8

Y

Y

Enter the date the claim
was taken. Format is
“CCYYMMDD”.

26

Effective
Date of
Claim

N

213

8

Y

Y

27

Program
Type

N

221

1

Y

Y

Enter effective date of
the claim. Correlates
with today’s date,
backdate reason, and
Liable State. Format is
“CCYYMMDD”.
Enter the program type:
1 = UI
5 = UCFE
7 = UCX

Y

Y

LAUS Program Manual 3-36

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

W/C IB

W/C COMM

28

Entitlement

N

222

1

Y

Y

Y

Y

29

SOC Code

N

223

4

3

3

3

3

30

Initial
Claim

N

227

1

Y

31

BYB

N

228

8

32

BYE

N

236

8

33

WBA

N

244

3

DESCRIPTION

Enter the entitlement
type:
0 = Regular
1 = Extended Benefits
(EB)
2 = Federal Benefit
Extension
3 = Additional Benefits
(AB)
Enter at least the first 3
digits of the Claimant’s
Occupational
Classification (SOC)
Code (left justified)
followed by a zero, or
enter the first 4 digits of
the SOC.
Enter Status of Claim:
1 = New
2 = Additional
3 = Transitional
Benefit Year Beginning
Date. Format is
“CCYYMMDD”.
Benefit Year Ending
Date. Format is
“CCYYMMDD
Weekly Benefit Amount
(Include Dependents
Allowance)

LAUS Program Manual 3-37

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

W/C IB

W/C COMM

DESCRIPTION

34

MBA

N

247

5

Maximum Benefit
Amount (Include
Dependents Allowance)

35

Base Period
Wages 1st Qtr
Base Period
Wages 2nd Qtr
Base Period
Wages 3rd Qtr
Base Period
Wages 4th Qtr
Base Period
Wages 5th Qtr
Base Period
Wages Total

N

252

7

Enter BP Wages for the
1st Qtr

N

259

7

Enter BP Wages for the
2nd Qtr

N

266

7

Enter BP Wages for the
3rd Qtr

N

273

7

Enter BP Wages for the
4th Qtr

N

280

7

Enter BP Wages for the
5th Qtr

N

287

8

Enter Total BP Wages
for all quarters

NAICS (Employer
with Most
Wages)

N

295

6

Enter at least the first
four digits of the North
American Industry
Classification System
(NAICS) Code (left
justified), followed by
“00”, for the employer
with which the claimant
had the most wages, or
enter the 6-digit code.

36

37

38

39

40

41

LAUS Program Manual 3-38

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

Last
Employer Name
Date
Employmen
t Began

A/N

301

30

N

331

8

Enter Date Employment
Began with Last
Employer. Format is
“CCYYMMDD

44

Date
Employmen
t Ended

N

339

8

Enter Date Employment
Ended with Last
Employer. Format is
“CCYYMMDD”.

45

NAICS –
(Last
Employer)

N

347

6

4

4

4

4

Enter at least the first
four digits of the North
American Industry
Classification System
(NAICS) Code (left
justified), followed by
“00”, for the claimant’s
Last Employer, or enter
the 6 digit code.

46

Last
Employer Ownership
Code

N

353

1

Y

Y

Y

Y

Valid entries are ‘1’
through ‘5’, default is
‘5’.
1 = Federal government
2 = State government
3 = Local government
4 = International or
Foreign
5 = Private

47

Separation

N

354

1

42

43

FIELD NAME

TIC

REOP/
TRAN

W/C IB

W/C COMM

DESCRIPTION

Enter name of Last
Employer.

Separation:
1 = Permanent
2 = Temporary

LAUS Program Manual 3-39

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

48

Recall Date

N

355

8

49

Union

A/N

363

1

50

US
Citizenship
Alien
Registration
Number
Week
Ending
Date
Earnings
During
Week
Claimed
Date First
Payment
Issued
Exhaustee

A/N

364

1

A/N

365

20

N

385

8

A/N

393

1

N

394

8

A/N

402

1

51

52

53

54

55

56

Weeks
Compensat
ed

N

403

2

57

$ Amount
of Benefits
Paid

N

405

7

TIC

Y

REOP/
TRAN

Y

W/C IB

W/C COMM

Y

Y

Y

Y

DESCRIPTION

Enter date claimant is to
return to work. If no
recall date, enter all
zeros, format is
“CCYYMMDD”.
Y = Yes
N = No
Y = Yes
N = No
Enter claimant’s Alien
Registration Number, if
applicable and available.
Week Ending Date of
week claimed, format is
“CCYYMMDD”.
X = Yes. Indicated that
claimant had earnings
during the week
claimed. Space = no
Enter the Date the First
Payment was Issued.
Format is “CCYYMMDD”.
X = Yes.
Complete only upon
exhaustion. Space = no
Enter the number of
weeks compensated
during the benefit year.
Enter the total amount
of benefits paid during
the benefit year.

LAUS Program Manual 3-40

APPENDIX TABLE 3-1
LIABLE/AGENT DATA TRANSFER RECORD
Required Fields by Record Type
Y = Required Field
N = Rule Number Listed at End of Layout

(Continued)
FLD
NBR

FIELD NAME

FIELD
TYPE

BEGIN
COLUMN

FIELD
LENGTH

TIC

REOP/
TRAN

W/C IB

W/C COMM

DESCRIPTION

58

Commuter
Identificatio
n Code

A/N

412

1

59

Reopen
Claim/Trans
fer of Claim

N

413

1

60

Ethnic

N

414

1

61

Filler

A/N

415

57

62

Record
Type

A/N

472

1

Y

Y

Y

Y

Required Entry to
indicate Type of Record
1 = TIC (Telephone
Initial Claim)
2 = Weeks Claimed
3 = Reopen/Transfer

63

Process
Date

N

473

8

Y

Y

Y

Y

Enter the Date the claim
was Processed. Format
is “CCYYMMDD”.

Y

Y

Y

Y

X = Yes. Complete to
identify claims filed by
commuters from
Residence State. Space
= no
1 = Reopen claim.
Complete when there is
a break in claim series
not caused by
employment.
2 = Transfer of claim.
Complete when there is
a change in the
Residence/Agent State
with no break in claim
series.
0 = Neither

Enter claimant’s Ethnic
group. Valid values
are:
1 = Hispanic or Latino
2 = Non-Hispanic or
Latino
3 = Not Available
(for future use)

LAUS Program Manual 3-41

APPENDIX TABLE 3-1

LIABLE/AGENT DATA TRANSFER RECORD
Rule
Number
1

2

3

4

5

6

Definition of Rule
On a Weeks Claimed (IB) – the Residence State FIPS can
be the same as the Liable State FIPS as long as the Agent
State FIPS is different.
Either Mailing Address or Field 23 (Residence County
FIPS) should be provided. For New England States,
either Mailing Address or Field 24 (Residence City/Town
FIPS) should be provided.
SOC should be provided if possible. If not provided, a
warning message will be returned. THE RECORD WILL BE
PROCESSED.
Last Employer – NAICS should be provided if possible. If
not provided, a warning message will be returned. THE
RECORD WILL BE PROCESSED.
On a Weeks Claimed – Commuter, the value in the
Residence State FIPS (field 20) will determine the
receiving state.
Residence Address will be completed when the Liable
State can provide a Residence Address that is different
from the Mailing Address.

LAUS Program Manual 3-42

APPENDIX TABLE 3-2
Agent Summary Report - Interstate Claims
01/03/00

LIABLE/AGENT DATA TRANSFER
PAGE 01
AGENT SUMMARY REPORT
STATE OF (Agent State Name)
FIPS ID IS: 00
AGENT STATE COUNTS FOR THE SAT. ENDING DATE 01/01/00

INITIAL CLAIMS
REG
EB
FSB
AB
TOTAL

UI
20
0
0
0
20

UCFE
0
0
0
0
0

UCX
0
0
0
0
0

TOTALS
20
0
0
0
20

REOPEN/TRANSFER
REG
EB
FSB
AB
TOTAL

UI
2
0
0
0
2

UCFE
0
0
0
0
0

UCX
0
0
0
0
0

TOTALS
2
0
0
0
2

UI
1358
0
0
0
1358

UCFE
24
0
0
0
24

UCX
11
0
0
0
11

TOTALS
1393
0
0
0
1393

WEEKS CLAIMED - IB
REG
EB
FSB
AB
TOTAL
01/03/00

LIABLE/AGENT DATA TRANSFER
PAGE 02
AGENT SUMMARY REPORT
STATE OF (Agent State Name)
FIPS ID IS: 00
AGENT STATE COUNTS FOR THE SATURDAY ENDING DATE 01/01/00
INITIAL CLAIMS
01 AL
ALABAMA
UI
UCFE
UCX
TOTALS
REG
0
0
0
0
EB
0
0
0
0
FSB
0
0
0
0
AB
0
0
0
0
TOTAL
0
0
0
0
02 AK
ALASKA
UI
UCFE
UCX
TOTALS
REG
1
0
0
1
EB
0
0
0
0
FSB
0
0
0
0
AB
0
0
0
0
TOTAL
1
0
0
1

LAUS Program Manual 3-43

APPENDIX TABLE 3-3
Agent/Residence Summary Report - Commuter Claims

01/03/00

LIABLE/AGENT DATA TRANSFER
PAGE 01
AGENT/RESIDENCE SUMMARY
REPORT
FIPS ID IS: 00
STATE OF (Agent State Name)
CURRENT MONTH = 12/1999 PREVIOUS MONTH = 11/1999
COMMUTER CURRENT
UI
UCFE
UCX
TOTALS
TOTALS
REG
184
2
0
186
EB
0
0
0
0
FSB
0
0
0
0
AB
0
0
0
0
TOTAL
184
2
0
186
COMMUTER PREVIOUS
TOTALS
REG
EB
FSB
AB
TOTAL

UI

UCFE

UCX

TOTALS

442
0
0
0
442

0
0
0
0
0

1
0
0
0
1

443
0
0
0
443

01/03/00

LIABLE/AGENT DATA TRANSFER
PAGE 02
AGENT/RESIDENCE SUMMARY
REPORT
FIPS ID IS: 00
STATE OF (Agent State Name)
COMMUTER CURRENT
CURRENT = 12/1999 PREVIOUS = 11/1999
01 AL ALABAMA
UI
UCFE
UCX
TOTALS
REG
0
0
0
0
EB
0
0
0
0
FSB
0
0
0
0
AB
0
0
0
0
TOTAL
0
0
0
0
02 AK ALASKA
UI
UCFE
UCX
TOTALS
REG
0
0
0
0
EB
0
0
0
0
FSB
0
0
0
0
AB
0
0
0
0
TOTAL
0
0
0
0

LAUS Program Manual 3-44

APPENDIX 3-4
Residency Assignment System User’s Guide
The Residency Adjustment System (RAS) is accessed using the EUSweb. State
LAUS technicians must have an EUSweb account. To obtain an account, State
technicians should contact their regional office. A server account consists of the
BLS assigned user ID along with a password and the necessary permissions to
access the BLS firewall and LAUS server.
States upload their inputs files through either the LAUS Program menu or, if the
State is participating in PROMIS, the PROMIS Program menu. RAS will
automatically process the files and return them to the appropriate State folder.
Email notification is automatically sent to State users to inform them that their
file has been processed and is ready to be downloaded.

LAUS Program Manual 3-45

Input Files
To use RAS, three files are needed for each claims file type to be processed; these
include the input file, the format file, and the definition file. States provide the
input file and the format file. The format file describes the layout of the input file
and must specifically identify the address fields. States are not limited to one file
type. For example, separate files can be processed for continued claims, initial
claims, LADT files or any other file a State may need to correct and geocode.
Corresponding format files must accompany each file type and individual RAS
job files must be setup for each file type.
Input File
The input file is a UI claimant record file that may consist of interstate claims,
interstate claims and/or commuter claims. The data for the input file is obtained
from the State’s UI branch. The file must be submitted in a consistent format for
periodic processing since any changes in the input file layout requires that the
format file and associated jobs files be modified to reflect the new layout.
At a minimum, records must include street address, city, & zip. Other
information, such as local office ID, claim dates, and claimant information, may
remain in the file. The conversion process will not affect these fields. The input
file must be an ASCII file in fixed record length format. As previously
mentioned, for the system to automatically process the input file, it must be
named appropriately with the State’s alpha abbreviation. The following are
accepted file names and their possible contents:
Standard Input File Names 
File Name

File Type

ST.txt generic file
STLADT.txt LADT file
STCOMM.txt commuter claims file
STCOM.txt commuter claims file variation
STHIS.txt generic history file
STANN.txt generic annual file
STPROMIS.txt weekly PROMIS input file
STLADTPROMIS.txt weekly PROMIS LADT input file
STINTRA.txt intrastate claims file
STREG.txt regular claims file
STCC.txt Continued claims file
STIC.txt initial claims file
ST = State alpha abbreviation

As long as the above criteria are met, the system can accept any file format. For
example, the complete LADT file can be submitted. Each record in the LADT file
contains 63 fields at a total of 473 bytes. The maximum number of fields per
record and the total number of bytes per record are limited together to 32,767
bytes.

LAUS Program Manual 3-46

RAS can also filter out unwanted records, such as claims with earnings and UCX
claims, prior to processing the input file. For large files, filtering out unwanted
records will speed up the RAS processing time.
In addition, large input files can be compressed before they are
uploaded to EUSweb using either WinZIP or PKZIP compression
utility software. Compressed files, or zipped files, use the same
naming conventions as above but must have the “ZIP” extension
instead of the “TXT” extension. The polling agent will automatically unzip any
files with the ZIP extension before sending them to RAS for processing. This
feature will significantly reduce the amount of uploading time for large input files
and eliminated EUSweb sessions from timing-out during the upload procedure.
Format File
The format file contains a list of each data field, its length and data type, and must
have the “FMT” extension. The file identifies the fields and their locations within
a record. Only the address fields (street, city, state, and ZIP code) need to be
specifically identified for RAS. It also identifies unnecessary fields that can be
ignored during processing or even removed before process begins. Data that the
State does not want to identify, such as Social Security numbers, should be
included in one large field along with other data. This way the positions of the
sensitive data are not identified.
The following is an example of a format file. In this case, the Social Security
number can be hidden in the “data” field which is 20 characters long.
 Format File Example 
data,20,c
address,30,c
city,19,c
state,2,c
zip,5,c
eor,2,b
The data type is “c” for most of the fields indicating that the field is a character
type. Only the “eor”, or end of record, is a “b” for binary type. The eor indicator
typically consists of two column lengths and is not visible unless using a text
editor, such as Word or TextPad, which allows the viewing of all spaces.
Although the eor is takes up two column places, it is usually identified as a single
symbol, such as in Word or in TextPad. The eor indicator can be viewed by
the clicking on the following button on the tool bar in Word or TextPad:
Output Files
States have much flexibility in how their output files are produced. The
output can be as simple as having the same file layout as the input file only with
corrections. It can also contain additional fields assigned by the system. RAS can
overwrite existing address fields with corrected data or the corrected data can be
written to new fields in any location in a record. FIPS for States, counties,
LAUS Program Manual 3-47

incorporated places, CDPs, MCDs and State specific codes can be also be written
to any location in the record or appended to the end of each record. Several
output files can be produced from one input file. Multiple output files can be
designated by claim types, such as regular and commuter claims, or other criteria
depending on the needs of the State.
Once a file has been processed, the output file will have the same name as the
input file but will include the word “out” (ex. STOUT.TXT). Multiple output
files would be named to reflect their contents.

LAUS Program Manual 3-48

4

Inputs to LAUS Estimation:
Establishment Data Sources

Introduction

T

here are two establishment-based data sources for employment estimates.
These are the Current Employment Statistics (CES) program and the
Quarterly Census of Employment and Wages Program (QCEW). The
next two sections provide an overview of these two programs.

The Current Employment Statistics Program
The Current Employment Statistics (CES) program is responsible for a FederalState cooperative monthly survey of 144,000 business establishments nationwide
that operates in all States, the District of Columbian, Puerto Rico, and the Virgin
Islands. These 144,000 businesses and government agencies represent
approximately 550,000 individual worksites. Each month, the survey provides
detailed industry data on employment, hours, and earnings of workers on nonfarm
payrolls for the nation, each State, and all metropolitan areas and areas in Puerto
Rico defined by the U.S. Office of Management and Budget.
The CES is an establishment survey that measures payroll jobs, unlike the Current
Population Survey (CPS) which is a household survey that measures employed
persons.
CES Concepts
Establishment. An establishment is defined in the CES as an economic unit, such
as a factory, mine, store, or office that produces goods or services. It generally is
at a single location and is engaged predominantly in one type of economic
activity. Where a single location encompasses two or more distinct activities,
these are treated as separate establishments if separate payroll records are
available and the various activities are classified under different industry codes.
LAUS Program Manual 4-1

Employment. Employment data refer to persons employed full- or part-time in
nonfarm establishments during a specified payroll period. Temporary employees
are included. In general, data refer to persons who worked during, or received pay
for, any part of the pay period that includes the 12th of the month. Government
employment covers only civilian employees; military personnel are excluded.
Employees of the Central Intelligence Agency, the National Security Agency, the
National Imagery and Mapping Agency, and the Defense Intelligence Agency
also are excluded. Persons are considered employed if they receive pay for any
part of the specified pay period, but they are not considered employed if they
receive no pay at all for the pay period. Therefore, persons 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 the rest of the period are counted as employed. Not
counted as employed 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 to work
during the pay period.
Since proprietors, the self-employed, and unpaid family workers do not have the
status of paid employees, they are not included. Also excluded from the employed
are farm workers and domestic workers in households. Salaried officers of
corporations are included.
CES Estimation
The estimation methodology for the CES program combines annual benchmarks
from the Quarterly Census of Employment and Wages program with monthly data
from a sample survey to produce estimates of employment, hours, and earnings.
All firms with 1,000 employees or more are asked to participate in the survey, as
is a sample of firms across all employment sizes. In 2014, the CES sample
consisted of about 150,000 businesses and government agencies that represented
approximately 550,000 individual worksites drawn from a sampling frame of UI
tax accounts. The sample rotation plan allows most firms to report for 4 years and
then be rotated out of the sample for a similar period. The sample frame is the
master list of establishments reporting to the Unemployment Insurance system.
Sample distribution is obtained by stratifying the universe of establishments for
each industry into employment-size classes. A total sample size sufficient to
produce adequate employment estimates is then determined and distributed
among the size classes in each industry based on the average employment per
establishment and the relative importance of each size class to its industry. Data
are then collected from the establishments surveyed on the report form BLS 790
or its electronic equivalent.
Employment estimates are made at the publication industry (or 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,
LAUS Program Manual 4-2

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 have 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.
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 based on 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 (non-aggregate) 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.
Employment estimates are controlled at what is termed the Estimation Super
Sector (ESS) cell level. ESS cells are those cells for which estimates are produced
that also have the special function of controlling Basic cells. The Basic cell can be
the same industry series as the ESS cell; or, if there are further industry breakouts
into detailed estimating cells, the ESS cell will control the sum of the Basic cells.
Estimates produced at the ESS level are generated using a larger sample size than
estimates produced at detailed estimating cells; therefore, the estimates produced
at the ESS level are statistically more robust and have less variance. Given these
more desirable properties at the ESS level, CES uses a top-down estimation
approach for published estimates.
CES State and Area estimation currently uses one of three separate estimators to
turn sample data into employment estimates for a given cell. Which estimator is
used for a cell depends on the characteristics of both that cell and the available
data.
1. The Robust Estimator is used to estimate employment at the ESS and nonESS levels when adequate sample exists.

LAUS Program Manual 4-3

2. The Small Domain Model (SDM) Estimator is used to estimate statewide
employment at the non-ESS level when inadequate sample data exists. It is
also used to estimate MSA employment at ESS and sub-ESS levels when
inadequate sample data exists. It uses the cell’s sample estimate as an input as
well as a model-based input of the QCEW trend. At the MSA level, the SDM
includes another input that uses the statewide link.
3. The Fay-Herriot (F-H) Estimator is used to estimate statewide employment at
the ESS level when inadequate sample data exists. It uses the cell’s robust
estimate as an input as well as a model-based input of the QCEW trend
combined with historical forecasts for all states and D.C. that have the CES
series.
Benchmarks
In order to control both sampling and nonsampling error, CES payroll
employment estimates are benchmarked annually to employment counts from a
census of the employer population. These counts are derived primarily from
employment data provided in unemployment insurance (UI) tax reports that
nearly all employers are required to file with State workforce agencies. The UI tax
reports are collected, reviewed, and edited by the staff of the BLS Quarterly
Census of Employment and Wages (QCEW) program. All employers covered by
UI laws are required to report employment and wage information to the
appropriate State workforce agency four times a year. Approximately 97 percent
of total nonfarm employment within the scope of the establishment survey is
covered by UI. A benchmark for the remaining 3 percent is constructed from
alternate sources, primarily records from the Railroad Retirement Board and
County Business Patterns. As part of the benchmark process for benchmark year
2013, QCEW-derived employment counts replaced CES payroll employment
estimates for all 50 States and the District of Columbia, Puerto Rico, the U.S.
Virgin Islands, and about 400 metropolitan areas and divisions for the period of
April 2012 to September 2013.
UI tax reports are not collected on a timely enough basis to allow for replacement
of CES payroll estimates for the fourth quarter, October 2013 to December 2013.
For this period, estimates based on existing sample information are revised using
the new series level from census-derived employment counts and updated
business birth/death factors.
Reliability of Estimates
Although the relatively large size of the CES sample assures a high degree of
accuracy, the estimates derived from it may differ from the figures that would be
obtained if it were possible to take a complete census using the same procedures.
Although the estimates are adjusted annually to new benchmarks, estimates
LAUS Program Manual 4-4

subsequent to the benchmark month have several potential sources of error. The
amount added each month for new establishments, for example, may be too high
or too low. Changes in the industrial classification of establishments that result
from changes in their product or activity between benchmark months are not
reflected. In addition, small sampling and response errors may accumulate over
several months as a result of the link relative technique of estimation used
between benchmarks.

LAUS Program Manual 4-5

The Quarterly Census of Employment and Wages Program
Background
The Quarterly Census of Employment and Wages (QCEW) program is a
cooperative endeavor of BLS and the workforce agencies of the 50 States, the
District of Columbia, Puerto Rico, and the Virgin Islands. Using data submitted
quarterly by the agencies, BLS summarizes employment and wage data for
workers covered by State unemployment insurance (UI) laws and for civilian
workers covered by the program of Unemployment Compensation for Federal
Employees (UCFE).
The QCEW program is a comprehensive source of employment and wage data by
industry at the national, State, and county levels. Unlike the CPS and CES
programs which are monthly sample surveys, the QCEW program is collected
quarterly and provides a virtual census of nonagricultural employees and their
wages. In addition, about half of all workers in agricultural industries are
covered.

Sources of Data
There are five sources of data for the QCEW program. They are initial status
reports, quarterly contribution reports, multiple worksite reports, Federal
Government reports, and annual refiling survey forms.
Initial Status Reports
Initial status reports are filed by new employers with the State UI tax unit to
initially register their business. These reports provide basic business identification
and classification information to establish a UI account. The employer’s liability
for UI taxes is determined from information provided in this report.
Quarterly Contributions Report (QCR)
QCRs are filed quarterly by all UI-liable employers to the State UI tax unit. These
reports provide the name and social security number of covered workers who
worked or received pay for the pay period which included the 12th of each month,
the total wages paid to covered workers, the portion of total wages subject to
unemployment insurance tax, and the employer contribution amount.
Multiple Worksite Report (MWR)
The MWRs are filed quarterly with the State QCEW unit by most employers with
more than one business establishment. The MWR provides establishment-level
employment and wages data not otherwise available on the QCR.
Federal Government Reports
These reports are filed quarterly by most federal government agencies to report
LAUS Program Manual 4-6

employment and wages data to the State QCEW unit, in accordance with the
Unemployment Compensation for Federal Employees (UCFE) program. Data for
non-defense federal agencies are provided to the State QCEW unit; information
for civilian employees of the Department of Defense is reported directly to BLSWashington.
Annual Refiling Survey Forms
UI-liable employers are surveyed by the State QCEW units periodically to verify
their location(s) and industry activity(ies). Employers are asked to verify physical
location, mailing address, and industry and ownership information and to provide
corrections if necessary

Data Compilation
State agencies code and summarize the raw data, check for missing information
and errors, and prepare imputations of data for delinquent reports. Each
establishment is classified by its industrial activity and then independently by one
of five ownership categories. (The ownership categories into which
establishments are classified are private industry, Federal government, State
government, and local government.) An establishment is an economic unit, such
as a factory, mine, or store, which produces goods or provides services. It usually
is at a single physical location and engaged in one, or predominantly one, type of
economic activity, for which a single industrial classification may be applied. The
North American Industry Classification System is used to classify the industry of
each establishment.
Employment data represent the number of workers on the payroll of covered
employers during the pay period including the 12th day of the month. Persons on
the payroll of more than one firm are counted in each firm. Workers are reported
in the State and county of the physical location of their job. Persons on paid sick
leave, paid holiday, paid vacation, and so on, are included, but those on leave
without pay for the entire payroll period, are excluded. The employment count
also excludes employees who earned no wages during the entire applicable period
because of work stoppages, temporary layoffs, illness, or unpaid vacations, and
employees who earned wages during the month but not during the applicable pay
period.
UI Coverage
Each State has determined its own laws regarding UI coverage, but they have
been greatly influenced by the federal government. The Federal Unemployment
Tax Act (FUTA) provides tax incentives that have ensured States’ conformity
with the minimum coverage standards set down in FUTA.
In general, a covered employer is defined under the FUTA as one who has a
quarterly payroll of $1500 in the calendar year or preceding calendar year, or one
LAUS Program Manual 4-7

worker in 20 weeks. While many States have chosen to expand coverage beyond
the FUTA standards, the notable exceptions and limitations are noted below.
Agriculture
For the majority of States, only employers with ten or more workers in twenty
weeks, or who paid $20,000 or more in wages in any quarter, are subject to
unemployment insurance laws. Farm owners/operators are excluded from
coverage in all states.
Domestic Service
Private households, social clubs, and college fraternities and sororities which
employ domestic help and pay wages of $1,000 or more in a quarter are subject to
unemployment insurance laws.
Nonprofit Organizations
Coverage is required for nonprofit organizations with four or more employees in
20 weeks. Almost half of the States, however, have elected more expansive
coverage, typically covering any organization with even one employee in twenty
weeks. Ministers employed by religious organizations to perform ministerial
duties are excluded from nonprofit coverage.
Self-employed Individuals and Unpaid Family Members
As defined by the unemployment insurance laws, employment is the hiring of
workers by others for wages. Self-employed individuals are therefore excluded,
except in California, where they may elect to pay contributions for self-coverage.
Relatives are not covered unless they receive pay from the official business
payroll. However, the employment of minors by their parents, or parents by their
children, is excluded.
Railroads
Interstate railroad workers are covered by the Railroad Unemployment Insurance
Act administered by the Railroad Retirement Board and thus are not included in
the QCEW data. Workers on intrastate and scenic railroads may be covered for UI
and included in the QCEW data.
State and Local Government Elected Officials and Others
All State and local government employees are covered under State UI laws
with the exception of elected officials, members of the judiciary, State
national and air national guardsmen, temporary emergency employees, and
policy and advisory positions.
Student Workers at Universities, Interns and Student Nurses
College and university students employed by the school at which they are
LAUS Program Manual 4-8

enrolled, such as work-study students, are excluded from coverage. Many States
also exclude the spouses of students who work at the university if the employment
is part of a program to provide financial assistance to the student. Student nurses
employed by hospitals as part of a training program are not covered. Similarly,
medical school graduates working as interns in hospitals are excluded from
coverage.
Armed Forces
Military personnel are excluded from State unemployment insurance coverage.
They are covered under a separate program, Unemployment Compensation for
Ex-Servicemen, but are not included in QCEW data. Civilian defense workers,
however, and all other federal employees covered under the Unemployment
Compensation for Federal Employees (UCFE) program are part of the data
reported to the QCEW program.
Agents on Commission
Insurance and real estate agents who are paid only by commission are
excluded from coverage in almost all of the States.
Earnings Data
Total wages, for purposes of the UI quarterly reports submitted by employers,
include gross wages and salaries, bonuses, tips and other gratuities, and the value
of meals and lodging, where supplied. In a majority of the States, employer
contributions to certain deferred compensation plans, such as 401(k) plans, are
included in total wages. Total wages, however, do not include employer
contributions to Old-Age, Survivors’, and Disability (OASDI); health insurance;
unemployment insurance; workers’ compensation; and private pension and
welfare funds.
Uses
The QCEW data serve as the basic source of benchmark information for
employment by industry and by size of establishment in the Current Employment
Statistics program. QCEW also serves as a national sampling frame for
establishment surveys by the Producer Price Index, Occupational Employment
Statistics, and other compensation programs.

LAUS Program Manual 4-9

Differences: Establishment Data Sources versus the CPS
The household and establishment data complement one another, each providing
information that the other cannot supply. Population characteristics, for
example, are obtained only from the CPS, whereas detailed industrial
classifications are much more reliably derived from establishment reports.
Certain differences can be accounted for, but others cannot. It is useful to be
aware of the CPS/CES/QCEW differences for estimation and analysis purposes.
Some of the important differences are discussed below.
Place of Work vs. Place of Residence. CES and QCEW data are produced
according to the location of the establishment; CPS data provide residency-based
employment estimates.
Jobs versus Employed People: CES and QCEW develop estimates of jobs. The
CPS estimates employed individuals. Workers holding more than one job may be
included more than once in the CES and QCEW employment counts since they
may appear on more than one payroll record or contribution report. Employed
persons in the CPS are counted only once even if they hold multiple jobs.
Reference Period Differences. The reference period for the CPS is the calendar
week including the 12th of the month, except in December when it is often the
week including the 5th. The reference period for CES and QCEW is the payroll
period including the 12th of each month, which could be weekly, biweekly, semimonthly, or monthly.
Employment Coverage Differences. The CPS definition of employment is total
employment, comprising wage and salary workers (including domestics and other
private household workers), self-employed persons, and unpaid workers who
worked 15 hours or more during the reference week in family-operated
enterprises. Employment in both agricultural and nonagricultural industries is
included. The CES and QCEW definitions reflect nonfarm wage and salary
employment and do not include self-employed and unpaid family workers.
QCEW includes some, but not all, domestics in private households and
agricultural workers, whereas these categories of worker are out-of-scope for
CES.
CES and QCEW estimates include 14- and 15-year olds while the universe for the
CPS is limited to 16 years of age and older. CES does not cover workers who are
on unpaid absence for the duration of the pay period. These workers may be
considered employed in the CPS depending on their job attachment as determined
in the course of the interview. Workers who are on strike for the entire pay period
of the establishment are not included in the CES and QCEW estimates, but are
considered employed in the CPS.

LAUS Program Manual 4-10

5

Inputs to LAUS Estimation:
Census Data

Introduction

T

he Census of Population and Housing, conducted every ten years by the
Bureau of the Census, is primarily intended to provide the population
counts necessary for apportionment of seats to the U.S. House of
Representatives and for determining legislative district boundaries.
Post-censal population estimates are used in the State CPS
estimation methodology and for estimating certain
unemployment components in the handbook methodology.
Official Census Population Estimates Program (PEP) estimates
are used to correct for discrepancies in ACS population estimates
at smaller geographies. Ongoing population estimation is
conducted by the Bureau of the Census through a Federal/State
cooperative program. Statewide population estimates are produced annually for
the United States, counties, and sub-county areas. Data are additive to the next
level of geography, i.e., the State is the sum of its counties. Except for the
decennial census year, population estimates pertain to July 1 of the reference year.
Prior to the 2010 census, the questionnaire came in two forms: the short form,
which includes questions found on every form (100-percent questions) and the
long form, which also includes sample questions. The majority of individuals
received a short form where questions regarding household relationships, sex,
race, age, marital status, Hispanic origin and housing are asked. Approximately
one-sixth of the population received the long form where sample questions
include the following topics: (1) social characteristics such as education, place of
birth, ancestry, disability, and veteran status; (2) economic characteristics such as
labor force status, occupation, industry, class of worker, place of work, work
experience, and income; and (3) more detailed housing questions.

LAUS Program Manual 5-1

The 2010 census was limited to the short-form questionnaire which was sent to all
households in the United States and Puerto Rico. Information previously gathered
by the long-form questionnaire is now provided by the Census Bureau’s
American Community Survey (ACS), which is collected in a sample of
approximately 3 million households annually. The ACS, released annually, has
become a primary source of economic and demographic data at detailed
geographic levels.
The LAUS program methodology uses ACS worker flows data for adjusting
establishment-based employment estimates to residency-based employment
estimates, for estimating certain employment components in the Handbook
methodology, and for disaggregating or apportioning handbook area estimates to
smaller areas.

The American Community Survey
The American Community Survey (ACS) is a relatively new survey conducted by
the U.S. Census Bureau. It uses a series of monthly samples to produce annually
updated estimates for the same small areas (census tracts and block groups)
formerly surveyed via the decennial census long-form sample. Initially, five years
of samples were required to produce these small-area data. The Census Bureau
released its first 5-year estimates in December 2010, and new small-area statistics
now are produced annually. The Census Bureau also produces 1-year data
products for larger geographic areas (i.e., areas with a population of at least
65,000). The ACS includes people living in both housing units (HUs) and group
quarters (GQs).
Topics covered by the ACS are virtually the same as those that were previously
covered by the census long-form sample data. Estimates are produced for
demographic characteristics, social characteristics, economic characteristics, and
housing characteristics.

ACS Sample Size
The full implementation of the ACS, which began in 2005, sampled
approximately 2.9 million housing unit addresses annually stateside. The 2011
ACS sampled approximately 3.3 million housing unit addresses—this
corresponds to an increase in the targeted annual sample size of 3.54 million
addresses that began with the June 2011 ACS sample. This increase in the
targeted annual sample size has continued since then, resulting in a 2013 ACS
sample of approximately 3.54 million housing unit addresses.

ACS “time-period” Estimates
In general, ACS data describe conditions over the time period during which the
Census Bureau collected the data (1 or 5 years); in contrast, Census 2000 data
described conditions around April 1, 2000. Thus, ACS data provide “time-period”

LAUS Program Manual 5-2

estimates, while prior long-form data provided approximate “point-in-time”
estimates. The time-period nature of ACS data has several important
consequences.
In the interpretation of 1-year estimates, ACS estimates describe the average
characteristics of an area over a 1-year time period during which the Census
Bureau collected the data. It is important to remember that while these estimates
represent the average characteristics over a single calendar year, the 1-year
estimates are not calculated as an average of 12 monthly values. Saying that ACS
estimates are “averages” seems simple and straightforward, but an “average” may
mean different things depending on the variable being studied. For example,
educational attainment is a population characteristic that changes slowly. In this
case, the ACS estimate is likely to provide a close approximation to conditions at
any one point during the year even if the Census Bureau collected the data over
the course of the year. Unemployment, on the other hand, fluctuates throughout
the year. The ACS unemployment estimate therefore may not be a close
approximation to the unemployment rate on any given day during the year. For 1year ACS data, the estimates do not represent conditions at any one point during a
year but do reflect data that is spread across the entire 1-year period.
The multiyear ACS estimates have the same interpretation as single-year
estimates, but, in these cases, the aggregated estimates represent information
collected over 5 years rather than 1 year. The longer data collection period means
that there is more time for conditions to change during the course of
measurement. In interpreting multiyear estimates, is should not be assumed that a
multiyear estimate represents conditions at the end of the multiyear period. Two
metropolitan areas may have the same 5-year average poverty rates but the
poverty rates of the two metropolitan areas may be substantially different on the
last day of the measurement period.
Comparisons involving single-year ACS estimates are straightforward because
they involve comparisons of data collected in two independent samples.
Comparisons involving multiyear ACS data are complicated by the fact that the
estimates being compared may be based on some of the same data. It is important
to distinguish between comparisons involving overlapping and nonoverlapping
samples. The 2005–2009 5-year estimates and the 2010–2014 5-year estimates are
said to be nonoverlapping because they use none of the same data. Comparisons
involving overlapping estimates depict changes only as calendar year data sets are
dropped and added to the sample. Any changes that may have occurred between
the beginning and the end of the survey period are muted by the use of the same
information (the overlapping samples) in both estimates. For example, changes
between the 2005–2009 5-year estimates and the 2006–2010 5-year estimates are
entirely driven by the changes between 2005 and 2010, because the 2006, 2007,
2008, and 2009 data are used in both estimates.
Overlapping estimates may give a sense of underlying trends but, because the
number of observations dropped and added between samples may be small, a

LAUS Program Manual 5-3

suspected trend may be the result of the deletion and addition of a relatively small
number of atypical cases. Nonoverlapping estimates will give the best indication
of whether and how conditions may have changed. Thus, rather than comparing
estimates from the 2005–2009 5-year data with the 2006–2010 5-year data, the
2005–2009 5-year estimates should be compared with the 2010–2014 5-year
estimates. Typical statistical tests cannot be applied to overlapping samples
because the overlap gives a false sense of the size of the samples being compared.

Nonsampling Error
The ACS data collection consists of 3 stages: 1) a mailed request to respond via
Internet and later followed by an option to complete a paper questionnaire and
return it by mail, if no response is received by Internet or mail then 2) computerassisted telephone interview (CATI) follow-up for nonresponse, and 3) computerassisted personal interview (CAPI) visit interviews on a subsample of the
remaining nonresponding sample addresses. The Census Bureau carries out ACS
nonresponse data collection with a permanent telephone and field interviewing
staff thoroughly familiar with the survey and its content.
Studies have indicated that the ACS procedures produce lower unit nonresponse
and item nonresponse than experienced by its predecessor survey, the decennial
long form. While the impact of sampling variation can be easily translated into
measures such as margins of error, it is difficult to actually measure how
improvements in unit and item nonresponse translate into more accurate data.

LAUS Program Manual 5-4

Differences: ACS versus CPS/LAUS Estimates
Data Collection and Publication
The CPS reference period is typically the week including the 12th of the month,
with interviews being conducted the following week (typically the week including
the 19th of the month). CPS data are produced and published monthly. Annual
average data are also developed at the end of the calendar year. The CPS uses a
fixed reference period, as compared to the ACS, where the reference period is the
week prior to when a respondent answers the survey. CPS interviews are
conducted in the course of a single designated week each month, whereas
respondents answer the ACS at times that vary throughout the month and year.
ACS respondents are initially contacted by mail and encouraged to complete the
survey via the Internet or to return a paper questionnaire. If they do not respond to
their survey within a month of receiving it, they are then contacted by phone.
Approximately 1 in 3 households that still do not provide answers are subsampled for an interviewer to contact them in person in the third month.
ACS responses can relate to any weekly period throughout the year and reflect
different economic events. Respondents can choose to delay completion of the
ACS form. ACS data are collected over a range of time periods. In the ACS, the
reference is to activity in the “last week” whenever the respondent fills out the
survey. In the CPS, the reference period is fixed. A varying reference week and
time of data collection could be particularly problematic for shorter, transitory
statuses or activities that could be influenced by seasonal variation.
Unemployment, for example, is a status that is subject to both seasonal and
cyclical variability.
The mode of collecting data also may affect the labor force estimates. All CPS
interviews are conducted through personal visits or telephone calls by Census
field representatives using laptop computers for data entry. ACS data are collected
primarily by internet and mail, with telephone and personal visit collection used
as follow-up to mail nonresponse. Data collected using paper forms do not have
interviewers assisting respondents in interpreting questions.
Both the ACS and CPS are sample surveys used to make estimates for a larger
population. Each person in the survey represents a larger number of similar
individuals in the population. To do this, each survey utilizes population estimates
produced by the Population Estimates Program at the Census Bureau. Each year,
the Population Estimates Program publishes population estimates by demographic
characteristics (age, sex, race, and Hispanic or Latino ethnicity) for the nation,
States and counties. The reference date for estimates is July 1st.

LAUS Program Manual 5-5

Labor Force Estimates
In 2013, the numbers of persons the ACS classified as “employed,”
“unemployed,” and “not in the labor force” for the nation were all higher than the
official CPS estimates. The ACS unemployment rate was 8.4 percent in 2013,
compared to the CPS 2013 annual average of 7.4 percent.
A number of factors may account for the difference in the estimates, including the
following: overall questionnaire differences, differing requirements in the two
surveys with regard to whether an individual is actively looking for work, and
differing reference periods and modes of collection.
The ACS questions relating to labor force activity are less detailed than those in
the CPS. For example, the ACS uses seven questions in determining labor force
status, while the CPS uses sixteen. There are more detailed, probing questions in
the CPS regarding employment status. In addition, the CPS information is always
collected by trained interviewers and never through internet or mail
questionnaires. The ACS instrument asks people if they are looking for work and
available to take a job if offered one, but does not ask about the nature of the job
search. The CPS questionnaire probes to see if people are actively looking for
work (interviewing, calling contacts, etc.) versus passively looking for work (for
example, looking at want ads in the paper). In the CPS, a person is unemployed
only if that person has actively searched for work.
Effective 2008, changes were made to the ACS questionnaire that modified and
improved existing questions for several subject areas. In particular, revised labor
force questions were introduced to better capture data on employment status. The
modifications had the impact of increasing the estimated number of employed
persons from the ACS relative to CPS and LAUS estimates.

LAUS Program Manual 5-6

Uses of ACS Data in LAUS
Uses of ACS in Labor Force Estimates
ACS non-agricultural wage and salary employment ratios derived from the
relevant ACS five-year dataset are used to disaggregate multicounty or
multi-MCD area estimates of establishment-based non-agricultural wage
and salary (NAWS) employment (M01) into the county components. M01
values at the county- or MCD-level are required for Handbook line 1
calculations. (See Chapter 7.)
In order to develop place-of-residence employment estimates for counties
and MCDs, ACS nonfarm employment levels and ACS worker flows data
are used in calculating residency adjustment factors that are applied to
monthly establishment-based employment estimates. (See Chapter 7
Section on Dynamic Residency Ratios.)
Area shares of Statewide ACS agricultural and all-other employment
(self-employed, unpaid family workers, and domestics in private
households) are used in developing current month handbook area
estimates of these employment components. (See Chapter 7.)
ACS employment estimates are used in the ACS-share disaggregation
method, which is used in conjunction with the claims-based
unemployment disaggregation method for county parts and cities. The use
of ACS data for disaggregating unemployment estimates is required when
UI claims data by county part or city of residence are not available. The
method uses ratios of ACS unemployment in subareas to the respective
larger area totals. (See Chapter 9.)
ACS worker flows data, which identify place of residence and place
of work in employment estimates, are used in the designation of
LMAs, including metropolitan areas, micropolitan areas, and small
labor market areas.

LAUS Program Manual 5-7

Uses of Population Data
Census population data are used to
resolve inconsistencies between ACS
population estimates and the annual
Census population estimates from the
Population Estimates Program (PEP).
ACS uses the PEP estimates for most
counties and Minor Civil Divisions in
their weighting methodology as controls,
however some smaller areas are
combined and controlled as one area. This can and does make some ACS
and PEP population estimates differ. The LAUS program resolves these
discrepancies by applying a ratio of the area’s Census annual population
estimate divided by the most recent 5-year ACS population estimate to
ACS estimates used in Line 2 and 3 employment shares and
disaggregation ratios. The impact of this population correction factor is
twofold:
1. ACS small area labor force estimates estimates are rebased to official
Census PEP estimates
2. 5-year ACS labor force estimates are updated to a current Census
population vintage year basis. This allows employment and
unemployment shares to keep pace with population change.
Decennial census population estimates for States, and the subsequent
postcensal estimates, serve as the population controls in CPS estimation.
In a ratio estimation procedure, known population totals are applied to
sample ratios to improve the accuracy of the sample-based estimates of
levels.
Age-specific population counts are used in the distribution of new entrant
unemployed and reentrant unemployed estimates to States. They are also
used in the disaggregation of handbook area entrants and reentrants to
cities and county parts.

LAUS Program Manual 5-8

ACS Employment/Unemployment Data
Data

Use

Total Employment
Total Unemployment
Employment:
Non-Agricultural Wage
and Salary

Agriculture
All-other
Commutation Data

Disaggregation of employment estimates
Disaggregation of unemployment estimates
Determination of appropriate weighting for
dynamic residency adjustment factors and to
disaggregate multicounty or multi-MCD area M01
inputs into single county and single MCD M01
inputs.
ACS agricultural employment shares of Statewide
CPS agricultural employment
ACS all-other employment shares of Statewide CPS
all-other employment
Definition of metropolitan, micropolitan, and small
labor market areas; also used in dynamic residency
ratio calculations

Population Data
16+ civilian, noninstitutional population
for States
Total population
Total population 16-19,
20+

CPS population controls
ACS population correction factors
Handbook area shares of Statewide CPS new and
reentrant unemployment;
Disaggregation of new and reentrant unemployment
from handbook areas to disaggregated areas

LAUS Program Manual 5-9

Post-Censal Population Estimation
Post-censal population estimates are used in the State CPS estimation
methodology and the ACS population correction factors applied to ACS estimates
used in Handbook lines 2 & 3 (all-other and agricultural employment,
respectively) and in the development of disaggregation ratios. Ongoing population
estimation is conducted by the Bureau of the Census through a Federal/State
cooperative program. Population estimates are produced annually for the United
States, states, and counties; sub-county estimates are produced biennially. Data
are additive to the next level of geography, i.e., the State is the sum of its
counties. Except for the decennial census year, population estimates pertain to
July 1 of the reference year.

State and County Total Resident Population
The goal of the state and county total population estimates process is to produce total
population estimates and estimates of the state population aged 18 and over for all states,
counties, and equivalents in the United States. Parishes in Louisiana, boroughs in Alaska,
and several independent cities (in Maryland, Missouri, Nevada, and Virginia) are treated
as counties. The process focuses on the development of estimates for counties (and
equivalents) only. State estimates exist only as a sum of the final estimates for counties.
The process involves estimating the population separately for ages under 18, 18 to 64,
and 65 and over. The three age groups are estimated for this process for two reasons.
First, different input data are used for domestic migration depending on if the population
is under age 65 (IRS tax exemptions) or 65 and over (Medicare enrollment). Second,
estimates of the state population aged 18 and over are produced and provided to the
Federal Election Commission.
Producing state and county total population estimates is similar to the production of
national estimates, as they are both based on the balancing equation. However, state and
county estimates are produced for annual July 1 dates, and they incorporate domestic
migration. Even though there are slight differences in the way we calculate the first three
months (April to July) from the estimates base (using only one quarter of a year of
migrants, for example), the process is very similar for all other points in the time series.
First the GQ population and “age” the population one year are subtracted in order to
produce an estimate of the household population at the start of each period. The aging
process takes the proportion of the previous vintage county population age 17 and 64,
applies that proportion to the current year, and moves that population into the next higher
age group (e.g., the estimated number of 64 year olds would “age” into the group aged 65
and over).
Net migration rates calculated from IRS and Medicare data are then applied to the aged
household population at the start of the period to create estimates of net domestic
migration. Then add net domestic migrants, add births (for the under 18 population),
subtract deaths, and add international migrants to produce an uncontrolled estimate of the
household population at the end of the period for each age group. The GQ population is
then added to create uncontrolled resident population estimates for each age group.
LAUS Program Manual 5-10

The next step in the process ensures consistency with the national estimates. First,
the calculated resident population numbers are controlled to equal the national
numbers by the three age groups. Second, the GQ change is added to the total
household domestic net migration estimate for each age group and control that
number to sum to zero at the national level by age group. Then final resident
population by age group is rounded and the remainder (usually very small) is
allocated to the largest population value in the country. Finally, the three age
groups are aggregated into total estimates for counties, and sum these estimates to
create final estimates for states.

LAUS Program Manual 5-11

6 Development of Statewide Estimates
Background

H

istorically, CPS samples have not been sufficiently large to produce
reliable monthly estimates directly from the survey for all States. As a
result, indirect methods have been used to estimate employment and
unemployment. As far back as 1960, Statewide estimates of employment and
unemployment were developed under uniform Federal procedures using the
Handbook method. With the introduction of CPS State estimates in the
1970s, a six-month moving average ratio adjustment to CPS levels
augmented the Handbook estimate. In the
late 1970s, the Levitan Commission was
established to review the measurement of
the labor force in the United States.
Among the recommendations made by the
Commission in its report of 1978 was that
BLS explore replacing the Handbook with
an econometric approach to subnational
estimation.

Building on work done by Mathematica Policy Research under contract to
BLS, preliminary models were developed in the mid-1980s. In order to
involve States directly in the research, the State Research Group, made up of
State Research Directors and BLS staff, was established in 1986 with the
support of the Interstate Conference of Employment Security Agencies
(presently called the National Association of State Work Force Agencies).
Regression and time-series techniques were employed, with the models

06/08/2009

LAUS Manual 6-1

extensively evaluated using empirical methods as well as recognized
statistical theory.
Modeling can address the small samples in each State which result in
unacceptably high variation in the monthly CPS estimates of State
employment and unemployment. To produce less variable labor force
estimates as well as produce more stable seasonally adjusted estimates, BLS
developed time-series models which ‘‘borrow strength’’ over time by using
historical series of sample observations for a given State to increase its
effective sample size. On average, the variance of month-to-month change in
the model estimates is about one third of the size of the CPS variance.
A type of regression model known as the Variable Coefficient Model (VCM)
best met the criteria. The VCM is so named because the coefficients in its
equations are allowed to vary over time to reflect structural changes in a
State’s data. The changing coefficients are estimated by the “forward filter”,
a widely used statistical technique that evaluates structural change against
sampling variability. The forward filter, also referred to as the Kalman Filter,
enables the VCM to handle the different relative accuracies that result when
an estimate draws upon data from several sources.
In 1988, a year of dual estimation of BLS and the States helped the states
make the transition from the Handbook to the VCM. In 1989, this new
method was implemented in 40 States and the District of Columbia. The
remaining States were using monthly CPS estimates of employment and
unemployment directly.
During the early 1990s, ongoing research at BLS brought about another
improved model that better dealt with error estimation and incorporated new
time-series variables. Known as the Signal-Plus-Noise model, it also uses
variable coefficients and the forward filter. The Signal-Plus-Noise model
was implemented in January 1994. In 1996, time-series modeling was
extended to the 11 more populous direct-use States because of reductions in
the size of the CPS sample.
The 2005 LAUS Redesign introduced a new third generation of LAUS
models. The objectives of the new generation models were to implement
direct model-based seasonal adjustment with reliability measures and to
improve the benchmarking procedure by incorporating real-time monthly
benchmarking. At the same time, 6 area models were introduced along with
corresponding Balance of State models.
Real-time benchmarking addressed a number of concerns with the prior
generation of LAUS models. It reduced annual revisions by incorporating
the CPS benchmark on a current basis. It eliminated prior model biases and
benchmarking issues. It ensured that national events and shocks to the
economy will be reflected in State estimates as they occur. It also eliminated
06/08/2009

LAUS Manual 6-2

the discrepancy between the sum-of-States estimates and the national notseasonally-adjusted totals. The following table illustrates differences
between the LAUS sum-of-States unemployment rates and the national CPS
rates for both seasonally-adjusted and not-seasonally-adjusted estimates in
the years prior to the introduction of the third generation LAUS models.
Difference Between LAUS sum-of-States and CPS national
unemployment rates, 1996-2004 (LAUS minus CPS)
Month

1996

1997

January
February
March
April
May
June
July
August
September
October
November
December

-0.2
-0.1
-0.2
-0.2
-0.2
0.0
-0.1
0.0
0.0
-0.1
-0.1
-0.1

-0.2
-0.2
-0.3
0.0
-0.1
-0.1
0.0
-0.1
0.0
0.1
0.1
-0.1

January
February
March
April
May
June
July
August
September
October
November
December

-0.1
-0.1
-0.1
-0.2
-0.2
0.0
-0.2
0.1
0.0
-0.1
-0.3
-0.2

-0.2
-0.3
-0.2
-0.2
0.0
-0.1
-0.1
0.0
0.0
0.0
0.1
-0.1

1998
1999
2000
2001
2002
2003
2004
Not Seasonally Adjusted
-0.1
0.0
-0.1
-0.2
-0.4
-0.3
-0.2
-0.1
-0.1
-0.1
-0.2
-0.3
-0.3
-0.1
-0.3
0.0
-0.3
-0.3
-0.4
-0.2
-0.3
0.1
-0.1
-0.1
-0.2
-0.3
0.2
-0.2
0.0
0.0
-0.2
0.0
-0.2
-0.3
-0.2
-0.1
-0.1
-0.1
-0.1
-0.3
-0.4
-0.2
-0.1
-0.2
-0.1
-0.1
-0.3
-0.2
-0.2
-0.2
-0.2
-0.2
-0.4
-0.3
-0.3
-0.2
-0.1
-0.1
0.0
-0.2
-0.1
-0.2
-0.1
-0.1
0.0
0.0
-0.3
-0.1
-0.2
-0.2
0.0
0.0
-0.1
-0.3
-0.4
-0.2
-0.2
0.0
0.0
-0.1
-0.4
-0.4
-0.1
-0.2
Seasonally Adjusted
-0.2
-0.1
-0.1
-0.2
-0.2
-0.1
-0.1
-0.1
-0.2
-0.2
-0.2
-0.1
-0.1
-0.1
-0.1
0.0
-0.1
-0.2
-0.1
0.0
-0.2
0.0
-0.2
-0.1
-0.2
-0.4
0.2
-0.3
0.0
0.0
-0.2
-0.1
-0.3
-0.3
-0.3
-0.1
-0.1
-0.1
-0.1
-0.4
-0.5
-0.2
-0.1
-0.2
-0.1
-0.1
-0.3
-0.3
-0.2
-0.1
-0.1
-0.1
-0.3
-0.2
-0.2
-0.1
0.0
-0.1
0.0
-0.3
-0.2
-0.3
-0.2
-0.1
0.0
0.0
-0.4
-0.2
-0.3
-0.3
0.0
-0.1
-0.1
-0.4
-0.5
-0.3
-0.2
-0.1
-0.1
-0.1
-0.5
-0.4
-0.1
-0.2

The new models also addressed consistency issues, ensuring that the sum of
State estimates equal that of the nation every month. As part of the real-time
benchmarking procedure, each month each State’s estimates were controlled
to a Census Division. There are 9 Census Divisions which were in turn
controlled to the national CPS. This controlling was done on a pro-rata basis,

06/08/2009

LAUS Manual 6-3

with all State estimates within a given Census Division, for example, being
adjusted by the same ratio.

In January 2010, BLS implemented smoothed seasonally adjusted (SSA)
estimates as the official seasonally adjusted series for States’ labor force data.
SSA estimates incorporate a long-run trend smoothing procedure, resulting in
estimates that are less volatile than those currently produced by the LAUS
estimation methodology. The use of the SSA methodology is effective in
reducing the number of spurious turning points in current estimates. More
importantly, SSA estimation can reduce revisions in historical estimates and
remove the potential disconnection between historically benchmarked and
current estimates.
The 2015 Redesign introduced a new fourth generation of LAUS models.
The Redesign changed how LAUS models incorporated time-series data
other than CPS data by changing to a regression model. In the new
regression models, time-series data used as covariates to the CPS have a fixed
relationship with the trend in an area’s estimates. This greatly reduces the
amount of computer resources required to produce estimates, especially when
revising historical estimates. Regression models are also more flexible,
opening up future possibilities for the modeling of covariate data.
The fourth generation models of the 2015 Redesign also contained other
features. They allow greater flexibility in the incorporation of outliers, such
as the option to have outlier adjustments spread over multiple months. They
also incorporate the process of real-time benchmarking, which had formerly
been an added step following model estimation. Model-based benchmarking
allows for the allocation of benchmark discrepancies according to the relative
volatility of individual series instead of according to a single pro-rata
adjustment.

06/08/2009

LAUS Manual 6-4

Model Structure
The model structure introduced in 2015 utilizes both univariate and regressor
modeling approaches. Univariate modeling is based only on the past values
of CPS unemployment or CPS employment and is utilized for Division
models. This approach combines a time-series model of the Signal and a
Noise model of the CPS survey. The survey estimates used in the models are
strengthened by the analysis of decades of monthly CPS sample data since
1976. For State and area estimates, a regressor modeling procedure is used.
Regressor modeling of the series depends on the past values of the CPS and a
related time series (payroll employment or UI claims) along with the
relationship between the CPS and the input series. Related time series like
payroll employment and UI claims help to interpret potentially spurious
movements in the CPS, but the model ultimately aims to produce what the
CPS would have estimated if it surveyed the entire population.
The State models incorporate five features that are tailored to the properties
of each State’s data series. These features include the smoothness of the
trend, the stability of seasonal patterns, survey error, the relationship of the
CPS trend to State input trends, and the presence and types of outliers.
Signal-Plus-Noise Approach
The State CPS estimates are broken out into
the Signal, which represents the true value
of the employed or the unemployed and the
Noise, which represents survey error
inherent in the CPS sampling procedures.
The observed CPS estimate consists of a true but unobserved labor force
value plus noise, which occurs because the estimates are derived from a
probability sample and not exhaustive measurement of the entire population.
CPS𝑡 = Signal𝑡 + Noise𝑡
Where:

Signal𝑡 = True Value
Noise𝑡 = Survey Error

The signal-plus-noise estimation models are based on a modeling approach
that accounts for and extracts the noise from the CPS time-series data, thus
providing a better estimate of the signal. An important component of the
noise in the CPS data is sampling error; its characteristics can be estimated
from survey-design information. Two other factors that account for the noise
are irregular movements in the data and occasional outliers. When there is a
change in the CPS level of employment or unemployment, that change is a
combination of the change in the true labor force signal and the change in the
06/08/2009

LAUS Manual 6-5

noise. The goal of the models is to isolate the signal from the noise to obtain
the best possible estimates of the true labor force values.
The models for employment and unemployment are a combination of two
processes: signal estimation and noise estimation. Noise estimation consists
of a model of sampling error predicted in the CPS. The design of the CPS
contains two predictable sources of error which noise estimation
incorporates. The first source is the sample size of the CPS for a given area,
which indicates the potential magnitude of error in a given month’s estimate.
The second source is the reuse of sample across time, linking the error in a
given month’s estimate to that in other months which share potentially
unrepresentative households with the given month’s sample. Noise
estimation responds to changes in CPS reliability as measured above. When
the CPS is more reliable, the CPS monthly estimate is given more weight in
determining the monthly signal estimate. When the CPS is less reliable, the
broader historical pattern of the CPS is given more weight in determining the
monthly signal estimate.
Signal estimation consists of a time-series model based on historical data
relationships, so a long historical CPS time series is required. While the
time-series model of the signal depends on past relationships, it does not
require that these relationships be fixed over time. A very important feature
of this model is a built-in self-tuning mechanism, known as the forward filter,
which automatically adjusts the regression coefficients and trend and
seasonal components to adapt to gradual structural changes as they occur.
Sudden, unpredictable changes in the time-series relationships are handled by
incorporating outlier effects into the model.
Noise
Accounting for CPS Sampling Error
There are two properties of the CPS, all controlled through the models, which
affect the time-series data: changing reliability and correlated sampling error.
Changing Reliability

Changing reliability is due to one or a combination of several factors. These
factors include survey redesigns after decennial censuses, sample-size
changes due to budget cuts or special supplementation, and variations in
labor force levels. Because of these factors, the CPS sampling error variance
is not fixed over time.
As the reliability of the CPS estimates changes, so do the weights used to
estimate the signal. The estimated signal is a combination of an estimate
based on the time-series model of the signal’s historical data and current CPS
estimates corrected by a model-based estimate of sampling error. The
06/08/2009

LAUS Manual 6-6

reliability of the CPS can change over the years. As it improves, less weight
is given to the time-series model and more weight to current CPS estimates.
The reverse is true for periods when the reliability weakens. Thus, the
estimated signal is a weighted average of a predicted signal based on
historical data and the current CPS estimate. This is represented by the
equation below:
Signal𝑡 = (1 − 𝑤𝑡 )Signal𝑡,𝑝𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 + 𝑤𝑡 CPS𝑡
Where:

Signal𝑡 = model estimate of the signal
Signal𝑡,𝑝𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 = model-based prediction of the signal
0 ≤ 𝑤𝑡 ≤ 1

The lower the reliability of the CPS, the less weight is placed on the current
CPS; the higher the reliability, the higher the weight.
Correlated Sampling Error

Because of the CPS 4-8-4 sample rotation method, there are significant
overlaps in the samples used by the CPS. (See Chapter 2.) Each month
three-fourths of the sample from the previous month is interviewed, oneeighth of the sample is interviewed for the first time, and one-eighth is
resuming interviews after being out of the sample for 8 months. Each month
one-half of the households from 12 months earlier are interviewed. The chart
below shows the proportion of the households in the current sample that were
also in the sample k months ago. For example, 75 percent of the households
in the sample this month were in sample last month, 50 percent were in two
months ago, etc. Note that samples from 4 to 8 months and over 15 months
apart have no households in common.

06/08/2009

LAUS Manual 6-7

Overlap of Identical Households
Percentage of Identical Households

80%

75%

70%
60%

50%

50%

50%
38%

40%
30%

25%

25%

20%

25%

13%

10%
0%

38%

1

2

3

0%

0%

0%

0%

0%

4

5

6

7

8

13%
0%

9

10

11

12

13

14

15 16 or
more

Months between Samples

The use of a rotation system requires the periodic replacement of the sample.
To cover a decade under the 4-8-4 scheme, 15 samples are needed. A key
feature of the replacement scheme is that successive samples are generated in
a dependent way. Once an initial sample of households is selected,
replacements are obtained from nearby addresses. For each original sample,
the 14 succeeding ones needed to cover the decade are usually taken from the
same neighborhood.
The overlap in the CPS sample is important because it introduces strong
autocorrelation in the sampling error. That is, the current value of the
sampling error (either an overestimate or an underestimate of the true value)
will depend on its own past values. For example, suppose the unemployment
rate for the sampled households in the current month is higher than the rate
for the entire population. Since 75 percent of these households will remain in
the sample next month, the unemployment rate is likely to be overestimated
again.
The extent of this autocorrelation depends not only on the overlap in the
sample but also on the stability of the labor force characteristic being
estimated. The overwhelming majority of workers spend most of their time
in the labor force as employed rather than unemployed. Accordingly, errors
in the employment estimates will be more strongly autocorrelated than in
unemployment estimates since employment is a more stable characteristic of
the households being sampled.
While CPS estimates of standard error have been routinely produced for State
CPS data, estimates of the error autocorrelations have not. Obtaining this
information is potentially very costly, involving complex calculations on
06/08/2009

LAUS Manual 6-8

huge micro data files. However, as part of BLS model research activities, a
method has been developed to estimate the autocorrelations that requires only
State CPS data for each rotation group.
The following graph presents the profile across time of a CPS error series for
a typical State for both employment and unemployment. The vertical axis
gives the weights that show how the effect of a CPS error occurring in a
given month is distributed over future months, and the horizontal axis
indicates the number of months following the occurrence of the error. For
the current month (zero months ahead), the weight equals one since the full
impact of the error is felt in the month that it occurs. A value of 0.6 for 1
month ahead, for example, indicates that 60 percent of the error in the current
month carries over into the next month’s estimate.

The strong autocorrelation in the CPS sampling error has important
consequences. First, sampling error will account for long-run movements in
the CPS. Ordinarily, we think of sampling error as having a transitory effect
on a series. If completely random, its effects would quickly average out.
This means the weights would drop to zero for all months following the
occurrence of the error.
Another important consequence of correlated sampling error is that these
errors are unlikely to average out over a 12-month period. If the error were
completely random, the number of overestimates and underestimates should
be about the same. However, because of the strong autocorrelation in the
error, if the CPS underestimates one year, it is likely to do so the next year.

06/08/2009

LAUS Manual 6-9

Signal
In addition to survey error, there are other sources of variation in the CPS
time series. These sources are identified as the seasonal, trend and irregular
components and are taken into account in the modeling procedure.
Based on the decomposition of a time series into the trend, seasonal,
irregular, and survey error components, the mode of seasonal adjustment may
be either additive or multiplicative. Formal tests are conducted to determine
the appropriate mode of adjustment.
The additive mode assumes that the absolute magnitudes of the components
of the series are independent of each other, which implies that the size of the
seasonal component is independent of the level of the series. In this case, the
seasonal factors represent positive or negative deviations from the original
series and are centered around zero. The seasonally adjusted series values are
computed by subtracting the corresponding seasonal factor from each
month’s original value.
In the multiplicative mode, the absolute magnitudes of the components of the
series are dependent on each other, which implies that the size of the seasonal
component increases and decreases with the level of the series. With this
mode, the monthly seasonal factors are ratios, with all positive values
centered around unity. The seasonally adjusted series values are computed
by dividing each month’s original value by the corresponding seasonal factor.
Time-Series Model of the CPS
CPS 𝑇 = {

𝑇𝑡 + 𝑆𝑡 + 𝐼𝑡 + 𝑒𝑡
𝑇𝑡 × 𝑆𝑡 × 𝐼𝑡 × 𝑒𝑡

Signal 𝑇 = {
Where:

CPS𝑡 − 𝑒𝑡
CPS𝑡 ÷ 𝑒𝑡

𝑇𝑡 = trend component
𝑆𝑡 = seasonal component
𝐼𝑡 = irregular component
𝑒𝑡 = survey error

Model Estimation
The first step of the estimation process is to correct the CPS estimate for
survey error.
CPS𝑡 − 𝑒𝑡∗ = Signal∗𝑡

06/08/2009

LAUS Manual 6-10

The second step is to seasonally adjust the error-corrected CPS. The seasonal
adjustment procedure is model-based. In a previous generation of models,
seasonal adjustment was performed independently of the model by the use of
ARIMA X-11 software. Once the seasonal factor is removed from the errorcorrected CPS, the remainder consists of the trend and the irregular
components.
Signal∗𝑡 − 𝑆𝑡∗ = 𝑇𝑡∗ + 𝐼𝑡∗
Example
The following example illustrates an additive time-series model of the CPS.
As mentioned above, the CPS is decomposed into trend, seasonal and
irregular components.
CPS𝑡 = 𝑇𝑡 + 𝑆𝑡 + 𝑒𝑡
Where:

𝑇𝑡 = global linear trend
𝑆𝑡 = fixed seasonal pattern
𝑒𝑡 = purely random (irregular)

The global linear trend model represents a linear relationship between the
dependent variable (CPSt) and t, where t indicates time. The magnitude and
direction of growth are fixed by the slope. The growth per period is
determined by the  value. The fixed intercept affects the level of the series.
The initial level is determined by  but has no effect on growth. The
smoothest possible trend would require that the growth lies on a straight line
centered through the series.
𝑇𝑡 = 𝛼 + 𝛽𝑡
𝑡 = 1, … 𝑁
The fixed seasonal model consists of 12 coefficients, one for each month of
the year. Each coefficient, or factor, measures the seasonal effect on the
series for a given month. Additive seasonal factors have positive and
negative values that indicate deviation above and below the trend level due to
seasonality. Summing all 12 months of seasonal factors will equal 0. The
fixed seasonal pattern repeats each year.
𝑆𝑚 = 𝑐𝑚
𝑚 = month index 1, 2, … , 12
𝑆1 + 𝑆2 + ⋯ + 𝑆12 = 0

06/08/2009

LAUS Manual 6-11

In this example the trend for time period t has an intercept of 10 and a slope
of 0.26.
𝑇𝑡∗ = 10 + 0.26𝑡
𝑆1∗ = 2.0, 𝑆2∗ = 3.5, 𝑆3∗ = 4.0, … , 𝑆12 = 0
First, the CPS is corrected for survey error. The survey error model uses an
autoregression approach with the current standard error as a weighted sum of
its previous values plus the current random error, vt. The coefficients are
provided from sample survey information.
𝑒𝑡 = 0.34𝑒𝑡−1 + 0.19𝑒𝑡−2 + 0.10𝑒𝑡−3 + 0.02𝑒𝑡−4 + 0.02𝑒𝑡−5 + 0.02𝑒𝑡−6,…
+ 𝑣𝑡
A regression equation is used to model survey error to account for the overlap
in CPS (autocorrelated error) and changes in reliability. The autocorrelated
component, et, is adjusted by a variance inflation factor (VIF). The VIF is
based on standard errors computed for the CPS.
𝑎𝑑𝑗

𝑒𝑡

= 𝑒𝑡 ∗ VIF𝑡

Removing the estimated error from the CPS yields a value for the signal
which equals the value for the signal computed from the trend and seasonal
components.
CPS𝑡 − 𝑒𝑡∗ = Signal∗𝑡
Signal∗𝑡 = (10 + 0.26𝑡 ) + 𝑆𝑚
𝑒𝑡∗ = CPS𝑡 − Signal∗𝑡
Next the error-corrected CPS is seasonally adjusted by removing the seasonal
component.
∗
Signal∗𝑡 − 𝑆𝑚
= 𝑇𝑡∗

Variable Regression Coefficients
The simple model in the above section may be generalized to handle real
series. Most series have changing trends and evolving seasonality. The trend
component in the example above cannot respond to a change in the direction
of the series. The seasonal component in the example above cannot respond
to changes in the seasonal pattern.
Using variable regression coefficients (VC) allows the coefficients to vary
over time so the model can adapt to changing patterns. In estimating the
coefficients, more recent observations receive more weight than earlier
06/08/2009

LAUS Manual 6-12

observations. Past data that are less relevant to current conditions are
discounted.
When the trend coefficients vary over time, the trend is able to adapt to
changing patterns. A poorly fitting fixed linear trend may miss important
turning points in the series. A variable slope and level associated with the
trend resolves this problem.
𝑇𝑡 = 𝛼𝑡 + 𝛽𝑡 𝑡
Similarly the seasonal component adapts to changing patterns and the
seasonal factor for a given month m will change over time. Seasonal factors
that are fixed may no longer reflect the current situation while seasonal
factors are made adaptive with the variable coefficient.
𝑆𝑚,𝑡 = 𝑐𝑚,𝑡
In addition, the VC is a self-tuning mechanism where the model adapts itself
without requiring special intervention. Each component has a
“hyperparameter” associated with it that determines how much it changes
over time. The hyperparameter is identified as i. If i = 0, then the
component is fixed. If i > 0, then the component changes continuously over
time. The hyperparameters are estimated from historical State data.
Component
Trend level (intercept)
Trend slope
Seasonal
Irregular

Fixed
level = 0
slope = 0
seasonal = 0
irregular = 0

Varying
level > 0
slope > 0
seasonal > 0
irregular > 0

Outliers
CPS data are occasionally affected by outliers. These outliers are CPS values
that are inconsistent with the expected behavior of either the signal or noise
component. There are two possible causes of an outlier: an unrepresentative
sample or a real non-repeatable event, such as bad weather or strikes.
Because these outliers represent sudden changes, they may cause special
problems for a model. In fact, we define an outlier as an observation that
breaks the pattern of behavior predicted by the model. It is not necessarily an
extreme value in the observed series. For example, a series may not change
much from one month to the next, but an outlier may have occurred if the
series normally has a large seasonal increase.
Even though there may be extreme observations in the CPS accompanied by
a few large prediction errors, it is not necessarily good practice to make
special adjustments to the model to fit those observations more closely. The
purpose of the model is to capture the normal time-series behavior of the
06/08/2009

LAUS Manual 6-13

signal. Thus, the model must be flexible enough to adapt to structural
changes in the signal, but, if the model is too flexible, it will fail to filter out
the noise. If there is prior information about the occurrence of an outlier,
then an adjustment may be justified. Otherwise, adjusting the model for
outliers is important only to the extent that they distort diagnostic testing,
cause bias in parameter estimates, or lead to a deterioration in current
performance.
Determining the type of outlier is crucial to deciding how to adjust the model
for its effects. Even though there are many complex patterns of outliers, the
three below tend to be the most common types of outliers that occur in timeseries data:




An additive outlier (AO) affects the series for only one month, such as
a sudden increase followed by a decrease.
A temporary change (TC) in the level of the series causes an abrupt
change in the series followed by a gradual return to its former level.
A permanent level shift (LS) refers to an abrupt shift that persists
indefinitely into the future, or until an offsetting shift in the opposite
direction occurs.

The outlier may be due to a real change in the labor force or result from the
measurement process, which includes sampling and other types of
measurement errors. The origin of the outlier determines whether it should
be included in the signal or the noise component. Ideally, this should be
resolved by seeking external information about the potential causes. In
practice, such information is rarely available. Since highly transitory outliers
in the CPS are more likely to be due to the measurement process than a real
economic event, the usual procedure is to assign these types of outliers (AO
and TC) to the noise component. On the other hand, a permanent shift in
level is considered a real effect and assigned to the trend component of the
signal. However, such identification requires that significant number of
months of data be available following the occurrence of the outlier in order to
identify the type of outlier that occurred. Therefore, outlier identification
cannot be made in current estimates. Models are monitored on a current
basis to detect the occurrence of outliers in the current year. Once enough
data become available to identify the nature of the outlier, its effects are
incorporated into the model specification as appropriate and implemented
during the annual re-estimation of the models.
Trend
Local Linear Trend Models
Univariate models of local linear trend are used to separately estimate the
trend component for CPS employment and unemployment, CES
establishment employment, and UI continued claimant unemployed. The
06/08/2009

LAUS Manual 6-14

local linear trend (Tt) of this type of model is comprised of a variable
coefficient trend where the intercept and the slope change over time.
𝑇𝑡 = 𝛼𝑡 + 𝛽𝑡−1
Where:

𝛼𝑡 = 𝑇𝑡−1 + ∇𝛼,𝑡
𝛽𝑡−1 = 𝛽𝑡−2 + ∇𝛽,𝑡−1
∇= per period change

A change in the local trend can result from a change in the intercept or a
change in the slope. When there is a change in the intercept, there will be a
shift in the level. The slope, on the other hand, changes the trend more
gradually.
The smoothness of trend is based on whether the intercept or the slope
accounts for most of the change. It the intercept is dominant, the trend
appears rough. If the slope is more important, then the trend looks smooth.
Based on this, the local linear trend can appear in three forms: smooth, rough
and general. The type of trend is determined from the data by empirically
estimating the variability in the intercept and slope (hyperparameters).
A smooth trend can result from the trend having a fixed intercept and a fixed
slope (global trend), or a fixed intercept and a changing slope. The trend line
shows continuous change but not abrupt shifts. Turning points in the series
are well defined with local peaks and troughs.
𝑇𝑡 = 𝛼𝑡 + 𝛽𝑡−1
Hyperparameters:
𝜎level = 0
𝜎slope > 0
A rough trend is caused by a changing intercept with no slope. All change is
due to shifts in the level. This gives the trend line a jagged look with many
small changes in direction. Occasional large shifts tend to be associated with
major business fluctuations.
𝑇𝑡 = 𝛼𝑡
Hyperparameters:
𝜎level > 0
The general form has shifting intercepts and slopes. It may approach
behavior of either the rough or smooth trend, depending on the relative size
of the change in the intercept and slope components.

06/08/2009

LAUS Manual 6-15

𝑇𝑡 = 𝛼𝑡 + 𝛽𝑡−1
Hyperparameters:
𝜎level > 0
𝜎slope > 0
Regressor Models
Trends of time series of the covariates to the CPS, CES establishment
employment and UI continued claimant unemployment, are estimated using
the local linear model and modeled seasonal adjustment methods described
above. The trends are then related to the CPS by regression to allow the
CPS’s covariate time series to contribute a part of the variation in their
respective CPS trends.
The local linear model approach takes steps to correct the CPS trend for
survey error. It is represented in the following equation.
𝑇CPS,𝑡 = 𝛼CPS,𝑡 + 𝛽CPS,𝑡
This creates what is now known as the CPS-specific trend, as it does not
reflect any consideration of its respective covariate. The regressor model,
which incorporates the State inputs previously modeled in the same fashion
as above, is represented in the following equation where “X” references the
UI or the CES.
∗
𝑇𝐶𝑃𝑆,𝑡 = 𝑇𝐶𝑃𝑆,𝑡
+ 𝛼𝑋 𝑇𝑋,𝑡

Where:

𝑇𝐶𝑃𝑆,𝑡 = total CPS trend
∗
𝑇𝐶𝑃𝑆,𝑡
= CPS − 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑡𝑟𝑒𝑛𝑑
𝛼𝑋 𝑇𝑋,𝑡 = component explained by covariate

The coefficient αX quantifies the degree to which the covariate’s estimated
trend influences the estimation of the CPS trend. It is intended to be fixed
throughout the entire time series. In practice, αX is fixed and constant
throughout the historical series at a single value estimated during the most
recent round of Annual Processing. In the months of current-year estimation,
the value is re-estimated during each round of monthly processing using the
entire historical series, though it will only impact the data being created
during processing until the next round of Annual Processing.
Area Models
In 2005, area models were introduced for the following areas:

06/08/2009

Chicago-Naperville-Arlington Heights, IL metropolitan division
LAUS Manual 6-16







Cleveland-Elyria, OH metropolitan statistical area
Detroit-Warren-Dearborn, MI metropolitan statistical area
Miami-Miami Beach-Kendall metropolitan division
New Orleans-Metairie-Kenner, LA metropolitan area
Seattle-Bellevue-Everett metropolitan division.

(Model-based estimation of the New Orleans-Metairie-Kenner, LA,
metropolitan area was suspended following Hurricane Katrina.)
Each of these area models is paired with a Balance of State (BOS) model.
The BOS is modeled directly rather than computing it as a residual of a State
model less the area model. This is done because modeling the BOS allows
the modeling of its separate error component. If the residual approach were
used, all the error would be allocated to the BOS. In addition, the BOS CPS
data are sometimes more reliable than the area data.
At the area level, covariate models are used. Both the unemployment the
employment models are additive. The area model and the BOS model for
each State are controlled pro-rata to that State’s model estimates.
Due to changes in the structure of the OMB-delineated geography of the
Chicago-Naperville-Arlington Heights, IL metropolitan division, modeled
estimates for this area begin in 1994, as opposed to 1990 as is conventionally
done with substate estimates.

Description of the Estimation Process
Overview
The LAUS program uses two approaches to estimation, real-time and
historical. Real-time estimation is a sequential process creating estimates one
month at a time, immediately after each new CPS estimate becomes
available. Historical estimation, on the other hand, is a batch process. Data
are accumulated over time and processed together at once.
Real-time estimation produces up-to-date estimates without delay. However,
there are some disadvantages. The model has less context with which to
evaluate potentially spurious movements in the time series. This renders the
trend component less smooth, since it is more likely to include error and
seasonality. Time and resource constraints limit how much both inputs and
outputs may be revised, so in practice we only revise the previous month
when generating a new month’s estimate.
The historical estimation process addresses the disadvantages associated with
real-time estimation. It better estimates error and evolving seasonal patterns
because it makes use of more contextual information, past and future; the
remaining trend estimate is clearer. The passage of time makes available
06/08/2009

LAUS Manual 6-17

more accurate unemployment insurance claimant and payroll employment
data. Outlier identification becomes vastly easier. The only drawback is that
the estimates are not timely.
Forward Filter
The real-time estimation procedure uses the forward filter algorithm. To
produce the current month’s estimate the forward filter requires only two
inputs. One is the prior month’s estimates of the signal and noise. The other
input is the current month’s CPS, CES, UI claims and population data. No
other historical data are needed.
The forward filter produces estimates at time t, taking into account
information available at this time. As new CPS data become available after t,
the estimate at t is not revised with new CPS data. States have the
opportunity to update their UI claims and CES data each month, but prior
month estimates are not updated with the current month’s data. Thus this
procedure only goes forward and does not look backward.
The forward filter provides estimates without delay. It is computationally
efficient as it requires no more work to process the last observation than it
does the first. In current practice, real-time estimation employing the forward
filter is carried out on both the most recent month for which CPS data is
available (generating a preliminary estimate) and the prior month (generating
a revised current-year estimate subject to future revision in Annual
Processing).
Model Re-Estimation
Historical estimation utilizes a model re-estimation algorithm to revise the
forward filter estimates to incorporate all data that become available after the
estimate for reference month. It accounts for all information available after
time t, i.e., over the whole available time series. Re-estimation is done once a
year and requires processing both current-year and historical data.
The re-estimation process uses all available data and thus is more accurate
than the forward filter procedure. It is less sensitive to the error present in the
CPS. As a result, it provides much smoother model components. In practice,
while the entire time series goes through model re-estimation, only several of
the most recent years of the time series will include substantial revisions to
inputs that therefore merit the publishing of newly revised estimates.
Real-Time Benchmarking
The purpose of benchmarking is to control for potential bias. Without
benchmarking, State models can be slow to adapt to national shocks.
Movements in the CPS time series at the level of the State are more likely to
be interpreted as statistical noise even when such movements may be borne
06/08/2009

LAUS Manual 6-18

out at the national level. To address these limitations, the present generation
of models incorporate monthly controls so that they sum to the national not
seasonally adjusted CPS. A much larger sample makes the national CPS a
more reliable benchmark. This constraint ensures that model employment
and unemployment estimates will adapt to national shocks without delay.
All benchmarking, whether in the current year or the historical series, is realtime benchmarking because each month’s estimate is controlled to its
respective estimate from the national CPS.
The procedure to benchmark estimates
to the national CPS is comprised of
three stages. In Stage 1 the model
estimates are produced for the nine
Census Divisions. As with the State
models, the Division model structure
is based on the decomposition of the
CPS time series into trend, seasonal,
error, and irregular components. No
covariate series are used. The
aggregated Division series are
constrained to sum to the national
CPS.

National CPS
(Control)

Division Model of
the Signal (Stage 1)
State Model of the
Signal (Stage 2)

In Stage 2 the State model estimates in
each Division are summed to the
Area Model of the
benchmarked Division model estimate
from Stage 1 (New York city, the Los
Signal (Stage 3)
Angeles-Long Beach-Glendale, CA
Metropolitan Division, and their respective balances of state are estimated as
States here). In Stage 3 the area model estimates are summed to the
benchmarked State estimate from Stage 2.
The benefits of real-time benchmarking to the national CPS are numerous. I
makes estimates consistent with reliable monthly national estimates. It
provides protection, particularly during real time estimation, from national
shocks to the economy such as recessions or catastrophic events like the
September 11 terrorist attacks. It provides more consistency between the
current-year estimates and historical estimates. Revisions are smaller when
transitioning from current-year estimates to historical estimates. However,
this procedure may introduce additional variability into the current-year
estimates due to fluctuations in the benchmark adjustment made each month.
The two approaches to benchmarking are external adjustments and internal
adjustments. External adjustments are made after estimation and are referred
to as pro-rata or ratio adjustment. This type is used in the real-time
benchmarking of the area models. Internal adjustment occurs during
06/08/2009

LAUS Manual 6-19

estimation and is the type currently used in the real-time benchmarking of
State and Division models. The latter method can produce better reliability
measures. In practice, the estimation process combines these approaches.
Using internal adjustment, Division model estimates of signal (i.e. not
seasonally adjusted estimates) are benchmarked to the national not seasonally
adjusted CPS. The nine Division models are combined into a “Super Model”
whose summed value is constrained to the national CPS estimate. Each
Division model is structured as follows:
Stage 1
′
′
Signal𝐷𝑖𝑣,𝑡 = 𝑤𝐷𝑖𝑣,𝑡
CPS𝐷𝑖𝑣,𝑡 + (1 − 𝑤𝐷𝑖𝑣,𝑡
)Signal𝑝𝑟𝑒𝑑
𝐷𝑖𝑣,𝑡 + 𝑏𝐷𝑖𝑣,𝑡 𝐵𝐷𝑖𝑓𝑓𝑈𝑆,𝑡

Where:

Signal𝐷𝑖𝑣,𝑡 = benchmarked Division signal estimate
Signal𝑝𝑟𝑒𝑑
𝐷𝑖𝑣,𝑡 = unbenchmarked Division signal prediction
CPS𝐷𝑖𝑣,𝑡 = current Division CPS estimate
𝑛𝐷𝑖𝑣

𝐵𝐷𝑖𝑓𝑓𝑈𝑆,𝑡 = CPS𝑈𝑆 − ∑ Signal𝑝𝑟𝑒𝑑
𝑖,𝑡
𝑛𝐷𝑖𝑣

𝑖=1

∑ 𝑏𝑖,𝑡 = 1
𝑖=1

′
𝑤𝐷𝑖𝑣,𝑡
+ 𝑏𝐷𝑖𝑣,𝑡 = 1

Each Division is estimated to have contributed (via bDiv,t) to the overall
benchmarking discrepancy BDiffUS,t based on the magnitude of that
Division’s CPS sampling error and the historically observed stability of its
trend and seasonal components. These factors together roughly correspond to
the relative sizes of each constituent Division’s labor force. In effect,
Divisions expected to have more reliable unbenchmarked estimates are
allocated a smaller proportion of the benchmarking discrepancy in each
month.
In the same manner as above, State model estimates are benchmarked to their
respective Division controls developed in the previous step. All States within
a Division are combined into a “Super Model” whose summed value is
constrained to the benchmarked Division estimate. Each State model is
structured as follows:
Stage 2
′
′
Signal𝑆𝑡,𝑡 = 𝑤𝑆𝑡,𝑡
CPS𝑆𝑡,𝑡 + (1 − 𝑤𝑆𝑡,𝑡
)Signal𝑝𝑟𝑒𝑑
𝑆𝑡,𝑡 + 𝑏𝑆𝑡,𝑡 𝐵𝐷𝑖𝑓𝑓𝑆𝑡,𝑡

Where:
06/08/2009

LAUS Manual 6-20

Signal𝑆𝑡,𝑡 = benchmarked State signal estimate
Signal𝑝𝑟𝑒𝑑
𝑆𝑡,𝑡 = unbenchmarked State signal prediction
CPS𝑆𝑡,𝑡 = current State CPS estimate
𝑛𝑆𝑡

𝐵𝐷𝑖𝑓𝑓𝑆𝑡,𝑡 = Signal𝐷𝑖𝑣,𝑡 − ∑ Signal𝑝𝑟𝑒𝑑
𝑖,𝑡
𝑖=1

𝑛𝑆𝑡

∑ 𝑏𝑖,𝑡 = 1
𝑖=1

′
+ 𝑏𝑆𝑡,𝑡 = 1
𝑤𝑆𝑡,𝑡

The final step, due to time and resource limitations, is to simply pro-rate the
area model estimates by controlling them to the benchmarked State estimate:
Stage 3
Signal𝐴𝑟𝑒𝑎,𝑡 = Model𝐴𝑟𝑒𝑎,𝑡 (

Signal𝑆𝑡,𝑡
)
𝑛𝐴𝑟𝑒𝑎
∑𝑖=1 Model𝑖,𝑡

In real-time benchmarking, the model estimate that is directly adjusted is the
not-seasonally adjusted estimate, the sum of the trend and seasonal
components. Implicitly, all the components are also ratio adjusted by the
same factor:
Adjusted trend𝑡 = 𝑘𝑡 trend𝑡
Adjusted seasonal𝑡 = 𝑘𝑡 seasonal𝑡
Where:

Signal𝑡
𝑘𝑡 = (
)
Signal𝑝𝑟𝑒𝑑
𝑡

Benchmarked model estimates may be somewhat more variable then the
original model estimates. Change over time in the original model estimates
are not exactly preserved due to fluctuation in the benchmark adjustment
factor from month-to-month. Only if the adjustment factors are equal to a
constant value over time will the proportionate change in the original model
estimates be exactly preserved:
Signal𝑡
𝑘𝑡 Signal𝑝𝑟𝑒𝑑
Signal𝑝𝑟𝑒𝑑
𝑡
𝑡
=
=
𝑝𝑟𝑒𝑑
Signal𝑡−1 𝑘𝑡−1 Signal
Signal𝑝𝑟𝑒𝑑
𝑡−1
𝑡−1
Where:

06/08/2009

𝑘𝑡 = 𝑘𝑡−1

LAUS Manual 6-21

For seasonally-adjusted estimates, an additional step is employed at this time
to reduce the volatility produced by real-time benchmarking. This step is
called "smoothing” and is discussed in the next section.
Smoothed Seasonally Adjusted Estimates
There are a number of sources of volatility in LAUS estimates. These
include sampling error in the CPS, real-time benchmarking, seasonality,
uncertainty at the end points of the series, frequent level shifts in the trend,
and real outliers in the current year (for example, Hurricane Katrina).
Volatility is a problem for model estimation when month-to-month change is
unexplained, is not related to predictable survey error or seasonal patterns,
lacks persistence, and is difficult to explain in terms of long-run movements.
Major sources of volatility in the models can be at least partially controlled
by the model. Normal survey error behavior is controlled with the surveyerror model. Outliers can be given special treatment. Seasonality, a
significant source of volatility, is removed with seasonal adjustment.
The principal volatility that we want to try to control is that which arises from
real-time monthly benchmarking. Apart from the inherent challenges of
accurately assigning a variable benchmarking discrepancy from to month,
benchmarking poses the additional challenge of constraining component
estimates to a control which retains its seasonality, thereby imposing some of
that seasonality via the process of real-time benchmarking. Because such
imposed seasonality is not a part of the component area’s series, it is not
captured through the model’s process of seasonal adjustment.
One approach to smoothing seasonally-adjusted benchmarked estimates is
through the use of moving averages, or filters. Moving averages “move”
through a time series from one period to the next by shifting the distribution
of time periods to be included. The center of this distribution is at time t, the
point being smoothed. The weights of all incorporated time periods add to
one. A simple moving average applies the same weight to all time periods.
A weighted moving average gives more weight to central observations,
making it more responsive to change. Asymmetric moving averages are
utilized for time periods with relatively fewer future observations (for
example, real-time estimates). Asymmetric moving averages necessarily
impose a lag on estimates.
The LAUS program uses a set of filters to smooth all of the points in the
series. Starting with a symmetric filter for points in the historical series,
asymmetric filters can then be derived that converge to the symmetric as
more data become available. This results in a “family of related filters”
consisting of a given symmetric filter and all of the necessary asymmetric
filters needed when there are not enough data points for the symmetric filter.

06/08/2009

LAUS Manual 6-22

The family of filters is known as a Trend-Cycle Cascade Filter (TCCF). The
TCCF consists of the weights of both the Henderson-13 trend filter and the
X-11 3x5 seasonal filter cascaded into a single, coherent filter whose weights
sum to one. The seasonal filter removes external seasonal patterns imposed
by real-time benchmarking, while the trend filter removes the balance of
month-to-month volatility not attributed to any particular source.
Annual Processing
The above benchmarking procedure does not eliminate the need for end-ofyear revisions; however, it does reduce the size of the revision compared to
the previous method. The smaller annual revisions to the real-time model
leads to less-over-the year distortion and facilitates analysis of the estimates
between historical and current year estimates.
Annual Processing involves entering revised UI claims, CES payroll
employment, and population estimates and performing model re-estimation.
The resulting estimates replace the real-time benchmarked estimates.

End-of-Year Processing
Revise population controls,
CPS, UI, & CES
Make historical estimates

Replace current-year estimates
Error Measures
In general, point estimates are never 100 percent
accurate. To convey the limitations of the data to our
data users it has been the Bureau of Labor Statistics’
policy to publish error measure if the methodology
permits. With the introduction of the new generation

06/08/2009

LAUS Manual 6-23

of models in 2005, we are now able to publish error measures.
There are two uses of error measures. One is reliability which gives us an
idea of how far are the estimates from the truth. The second is for analysis by
giving us an idea of what we can say about the truth.
Standard Error
The standard deviation of errors in the estimates gives a measure of the
dispersion of the error around a mean of zero. The larger the standard
deviation (Stder), the more likely an individual estimate is far from the true
values.
Example: A point estimate of 238,000 persons may have a Stder of 25,600
persons.
Coefficient of Variation
From the Stder other error measures that facilitate analysis can be computed.
The coefficient of variation (CV) is a common reliability measure that is
useful for comparing the estimates of different size or scale. The CV is
computed by dividing the standard error by the estimate.
CV =

Stder
25,600
=
= 0.11
Estimate 238,000

Confidence Intervals
The Stder is also used to construct confidence intervals. A confidence
interval for an estimate give us upper and lower limits around the estimate
where the true value likely to be located with a given level of certainty or
confidence.
Estimate ± 𝑘 ∗ Stder
If k = 1.96, the significance is at a 95 percent confidence level. If k = 1.645,
it is at a 90 percent level.
Significance
These error measures can be used to determine the difference between a State
estimate and the estimate for the nation or another State is statistically
significant. It is also used to reveal if the over-the-month change is
significant.
𝑧=

06/08/2009

Estimate − Mean
Stder

LAUS Manual 6-24

If z ≥ 1.96, then the difference is significant at the 95 percent level. If z ≥
1.645, then it is significant at the 90 percent level.
Error measures assist the analysis of month-to-month change, the differences
from the US and other State estimates and reliability.
Availability of Measures
Error measures for both smoothed-seasonally-adjusted and not-seasonallyadjusted estimates are available monthly in the STARS tables. Table 2
provides standard errors (Stder) for the point estimates, while Table 3
provides standard errors for over-the-year change, and Table 7 provides
standard errors for over-the-month change.

06/08/2009

LAUS Manual 6-25

LAUS Estimation:
Handbook Area Estimates

7
Introduction

n the late 1940s, when sub-national labor force estimation was first attempted,
employment and unemployment estimates were developed for large labor
market areas as well as States, underscoring the importance of substate labor
market information. In the 1960s, techniques for developing substate estimates
were published in the Handbook on Estimating Unemployment. Since then,
subsequent iterations of substate estimating procedures have been referred to as
the Handbook method and the areas for which these estimates are produced are
referred to as Handbook areas. (See Chapter 1 for more history on the Handbook
method.) Today, the LAUS program creates estimates for 4,756 Handbook areas,
the counties (Minor Civil Divisions in New England) that exhaust the geography
of all states, the District of Columbia, and Puerto Rico.

I

UI
Claims

Exhaustees

Non-Ag
Wage &
Salary

Entrants

All Other
Emp

Ag
Emp

Estimates for most Handbook areas are produced
independently by means of a building block
approach, which uses current unemployment
insurance (UI) data and current nonfarm
employment estimates as basic inputs. In
addition, components of the labor force not
covered by the basic source data are developed
using larger-area Current Population Survey and
American Community Survey estimate

relationships.
When the Handbook methodology was first introduced as a standard procedure
for sub-national labor force estimation in 1960, it was viewed as an attempt to
approximate the results of a CPS-type household survey, but without the
prohibitive cost of conducting such a survey. The Handbook method utilized a
system of estimates that was reflective of the labor market structure of the 1960’s
in terms of UI coverage. Over the years, refinements were made to the
components and basic input data to improve comparability and consistency within
the States and with the standard definitions of the labor force as embodied in the
CPS.
LAUS Program Manual 7-1

Today, the Handbook methodology consists of 14 line items that can be broken
out into employment and unemployment estimation procedures. This chapter will
provide details for each Handbook line and the associated inputs entered into the
LAUSToo system.

Handbook Line Items
Employment
Line

Description

1

Non-agricultural Wage & Salary Employment

2

All-other Employment

3

Agricultural Employment

4

Total Handbook Employment (lines 1 + 2 + 3)
Unemployment

Line

Description

5

UI Claims

6

UCFE Claims

7

Rail Road Claims

8

Total Claims (lines 5 + 6 + 7)

9

Unemployed Exhaustees

10

Non-covered Agricultural Unemployment

11

Unemployed excluding Entrants (lines 8 + 9 + 10)

13

Re-entrants

15

New Entrants

16

Total Unemployment (lines 11 + 13 + 15)

LAUS Program Manual 7-2

Additivity
Prior to 1977, the Handbook estimates were the final LAUS estimates for
Handbook areas. Beginning in that year, additivity of the substate Handbook
estimates to statewide estimates was introduced to address methodology issues
and Federal program allocation needs.
The sums of Handbook employment and unemployment for all Handbook areas in
a State tend to be lower than statewide estimates due to the greater difficulty in
obtaining some of the input data elements at the substate level. Forcing the
Handbook area estimates to sum to the statewide totals corrects for any
methodological deficiencies in a proportional manner, allowing for the complete,
to-the-dollar distribution of federal funds to areas when LAUS data are used in
the allocation algorithm.
Additivity is considered a separate methodological step that follows Handbook
estimation. (See Chapter 9 for more details.)

Geography of Handbook Estimation
Prior to 2015, Handbook estimation generally occurred at the Labor Market Area
(LMA) level. Exceptions to this general rule included Metropolitan Divisions that
subdivided Metropolitan Statistical Areas and Expanded Estimating Areas in New
England that comprised Metropolitan/Micropolitan NECTAs and one or more
isolated MCDs or an Adjacent Small LMA. Beginning in 2015, Handbook
estimation was shifted down to the county level outside of New England and to
the Minor Civil Division (MCD) level in New England. This change has several
benefits:


Elimination of Interstate Handbook Areas
Several transfers between States were eliminated. Prior to 2015, the States
with intrastate parts of interstate areas transferred continued claims and
final payments data for their respective parts to one State that was
identified as the controlling State for the whole interstate area. The
controlling State would then create Handbook estimates for the whole area
and for the disaggregated intrastate parts and would transfer the Handbook
part data back to the non-controlling States.
The old process had several drawbacks. It caused input corrections in one
State to ripple across neighboring States. It also included layered
disaggregation steps that made replication of calculations difficult. By
shifting Handbook calculation to the county/MCD level, the overall
process has become more streamlined with fewer cross-State exchanges
LAUS Program Manual 7-3

and many aspects of the methodology have become easier to apply and
understand.


Increased Specificity of Exhaustee Estimation (Handbook Line 9)
Previously, unemployed exhaustees were calculated at the LMA level
using final payments data that had been entered by county or New
England MCD. Exhaustees for each LMA were then distributed across the
counties or MCDs that compose the LMA in proportion to each county’s
or MCD’s share of the LMA’s continued claims. This approach was
problematic in that the distribution of continued claims could differ from
the distribution of final payments. In effect, exhaustees were shifted away
from the areas with the most final payments toward those with fewer final
payments within each LMA.
By moving Handbook calculations to the county or New England MCD
level, final payments data now are utilized in calculations at the lowest
geographic level possible and the disaggregation of exhaustees no longer
occurs.



Simplified Updates to OMB Geography
Before 2015, major updates occurred on a decennial basis for Combined
Statistical Areas, Metropolitan Statistical Areas, Metropolitan Divisions,
Micropolitan Statistical Areas, and the NECTA equivalents of these areas
in New England. Because many of these geographies were Handbook
areas, the impact on LAUS was substantial. Historic Handbook data had
to be realigned with new geography during annual processing. By
conducting Handbook estimation by county and New England MCD,
OMB geography updates are less disruptive to LAUS.

Weighed against the benefits above is one necessary new complication:


Separate Geography for Handbook Estimation and for Entering NAWS
Employment and Labor Disputants (M01 and M02)
Prior to 2015, establishment-based Non-agricultural Wage & Salary
(NAWS) employment data (M01) and labor disputant data (M02) were
entered by Handbook area (i.e., at the LMA level with some exceptions).
For some areas in select States, estimation and entry of M01 and M02
were shifted to the county level to mirror the change in geographic scope
for Handbook estimation. In general, though, M01 and M02 continue to
be entered by LMA.
To accommodate the discrepancy between the geography of these inputs
LAUS Program Manual 7-4

and the geography of Handbook estimation, a new disaggregation
component was added to the calculation of place-of-residence NAWS
employment (Handbook line 1).

LAUS Program Manual 7-5

Handbook Area Employment
Employment comprises all persons who did any work at all as paid employees,
worked in their own business, own profession, or on their own farm, or who
worked 15 hours or more as unpaid workers in an enterprise operated by a
member of the family. It also includes all those who were not working but who
had jobs or businesses from which they were temporarily absent because of
vacation, illness, bad weather, labor-management dispute, job training, child-care
problems, maternity or paternity leave, or other family or personal reasons,
whether or not they were paid for the time off or were seeking other jobs. (See
Chapter 2 for more details.)
The Handbook method decomposes employment into three subcategories. The
following table provides a brief summary of the employment inputs entered into
LAUSToo for each Handbook line. The sections following the table discuss each
item in greater detail.
Handbook Employment
Line

Line Description

1

Non-agricultural
Wage and Salary
(NAWS)
Employment

+

+

2

3

All-other
Employment

Agricultural
Employment

Input description
(LAUSToo variable ID)

Dynamic Residency Ratios
(DRRs)

BLS

ACS NAWS Employment
(F01)

BLS

Establishment-based NAWS
(M01)

State

Labor-management Disputants
(M02)

State

Statewide All-other Employment
(SAO)

BLS

ACS All-other Employment
(F02)

BLS

Statewide Agricultural
Employment (SAG)

BLS

ACS Agricultural Employment
(F03)
=

4

Input Source

BLS

Total Handbook Employment

LAUS Program Manual 7-6

Non-agricultural Wage and Salary Employment
(Handbook line 1)
The Handbook calculation of residency-adjusted non-agricultural wage and salary
employment begins with input data that pertain to jobs by place of work
(establishment data) rather than employed people by place of residence
(household data). The conceptual differences between (1) the establishmentbased inputs entered into LAUSToo and (2) the desired household-based output
require calculations to adjust the inputs to conform to CPS concepts. The
Handbook line 1 calculations apply Dynamic Residency-adjustment Ratios
(DRRs) to the establishment-based input data to bridge the conceptual gap.
The States provide two inputs that are entered into LAUSToo each month for the
line 1 calculations:


Establishment-based Non-agricultural Wage and Salary Employment
(NAWS)
o Also referred to as “place-of-work NAWS” or “pre-adjusted
NAWS”
o LAUSToo variable ID M01



Labor-Management Disputants
o Also referred to as workers involved in “work stoppages”
o LAUSToo variable ID M02

BLS provides two inputs that are entered into LAUSToo for the line 1
calculations:



Dynamic Residency-adjustment Ratios (DRRs)
Population-adjusted ACS NAWS Employment (F01)

The following sections will provide details regarding the sources and
development of these inputs.

Handbook line 1 input: Establishment-based Non-agricultural
Wage and Salary (NAWS) Employment (M01)
For most States, there is no single source of M01 data. Obtaining the M01 inputs
for all M01 input areas in a State usually requires the use of various data sources.
The principal source is the Current Employment Statistics (CES) survey monthly
estimates for over 400 Metropolitan Statistical Areas and Metropolitan Divisions,
which are the required inputs for areas where CES estimates are produced. For
those areas that are not within the CES program’s scope, a sample-based
LAUS Program Manual 7-7

employment series developed under State auspices is the next best data source. If
such a series is not available, a number of ways to produce the input can be used.
Details for the most commonly used data sources and procedures are provided
below.
Current Employment Statistics (CES) data
The CES survey, which is also referred to as the “payroll survey” or
“establishment survey,” yields employment estimates for all metropolitan areas
and most metropolitan divisions. It is the principle source of the M01 input and
sets the conceptual standard for M01 values derived from other sources. The CES
not-seasonally-adjusted estimate of total nonfarm wage and salary employment
should be used for LAUS estimation wherever it is available. No adjustment of
CES data is necessary to conform to the definition of the LAUS M01 input.
The CES program conducts monthly revisions (data for the prior month are
revised with the release of the current month) and annual benchmarking revisions
(data for the latest 18 months are revised at the end of each year). The LAUS
program incorporates these revisions during monthly and annual processing,
respectively. Historic CES corrections that extend beyond the standard two-year
LAUS annual processing period are incorporated if substantially large.
Quarterly Census of Employment and Wages (QCEW) data
QCEW data are often used to develop M01 inputs for Non-CES Areas. The
QCEW program publishes counts of establishments, jobs and aggregate wage data
that are derived from quarterly tax reports. The reports are submitted to State
employment security agencies by employers that are subject to either (1) State UI
laws or (2) the Unemployment Compensation for Federal Employees (UCFE)
program. Each quarter, QCEW releases monthly data. Data for any given year
are not final until data for the first quarter of the following year are released.
LAUS estimation generally occurs with only a one-month lag from the reference
period. In contrast, the volume of data processed by the QCEW program requires
a longer lag of 6 months between the reference quarter and the release of data. To
reconcile the timely needs of the LAUS program with the longer lag of the
QCEW program, extrapolation is used for monthly processing. During annual
processing, actual QCEW data replace extrapolated data.
Adjustments made to both extrapolations of QCEW data and actual QCEW data
are discussed below, followed by details of the extrapolation process.

LAUS Program Manual 7-8

Adjustments to QCEW data
Several steps are necessary to adjust QCEW job counts to conform to the
definition of the LAUS M01 input. Essentially, these adjustments bring QCEW
data into alignment with the CES data used for metropolitan areas.
1. Start with the QCEW total job count for the M01 input area.
2. Subtract agricultural jobs (All NAICS 11 codes except for 1133).
o These are accounted for in the Agricultural Employment
estimate (line 3).
3. Subtract private household workers (NAICS 814).
o These are accounted for in the All-other Employment estimate
(line 2).
4. Add Noncovered Employment (NCE).
o This includes various categories of jobs that are not covered by
unemployment insurance laws, but which are included in CES
estimates. These include:




College students working for their schools,
Commissioned insurance agents, and
Religious employees.

o Most States have a periodic survey that estimates NCE at the
State and metropolitan area level. The LAUS technician
should consult with the CES unit in their State to obtain NCE
estimates for Handbook areas.
5. Add estimates of railroad employment.
o These estimates are provided to States by the BLS national
office.
Extrapolation of QCEW data
Due to the time lag associated with publication of QCEW data, extrapolation is
necessary for LAUS monthly processing. A possible approach follows:
Extrapolation using Area-based Change Factors


Historic over-the-month changes for each reference month are
applied to the prior month’s estimate.



The historic monthly changes can be those for the prior year or
averages of those for several prior years.



For example, to estimate March 2010, take the over-the-month
QCEW employment change from February 2009 to March
2009 and apply that factor to the February 2010 estimate.
LAUS Program Manual 7-9

Small Area Employment Estimates (SAEE)
BLS Small Area Employment Estimates are nine-month forecasts of QCEW data
for Handbook areas for all States (except Maine). The forecast program utilizes
historical monthly QCEW data for each Handbook area in a State. Once the most
recent quarter of QCEW data are available to national office staff, ARIMA
software is run producing employment forecasts for each Handbook area.
Forecasted employment counts are aggregated to the M01 input area level and
provided to some States via EUSWeb. States need to add NCE and railroad
estimates to these forecasts before utilizing them as inputs to LAUS employment
estimation.
BLS currently maintains SAEE models for every county and New England MCD
outside of Maine. However, BLS provides data to states by M01 area, which
means aggregations of forecasts are provided for some areas rather than just the
components. These data include SAEE forecasts for CES areas, but these values
are not to be used in official estimation as current CES estimates are the required
input in monthly production for CES areas.
As of the 2015 LAUS program redesign, the total number of SAEE models is
5,195 (an additional 633 will be available once ME is completed). BLS provides
forecasts for 1,944 M01 areas, plus an additional 651 CES areas (this would
increase by 26 M01 and 7 CES areas when ME is completed).
Although not every State requests SAEE forecasts, they are produced and
distributed to all States. To use SAEE forecasts directly in monthly production,
States need to request approval through their Regional Office. However, formal
approval is not required for States who wish to request the forecasts as a review
tool in developing their own M01 inputs.
Small Domain Estimators
Illinois uses the National Opinion Research Center (NORC) Small Domain
Estimator model to produce non-agricultural employment by industry for their
non-CES areas. The model utilizes CES sample, QCEW employment, NCE
estimates, allocation of employment for statewide reporters, and economic events
not captured in the CES sample. At the end of each year, the model estimates are
benchmarked to QCEW employment.
Pennsylvania uses a Small Domain Estimator that is based on the inverse
relationship between unemployment insurance claims and employment—when
workers get laid off, claims go up and employment goes down; when workers
return to work, claims go down and employment goes up.

LAUS Program Manual 7-10

CES ACESweb and Composite Quota Estimator (CQE)
Some States utilize features within the CES ACESweb estimation production
system to produce employment estimates for areas not included in the CES
program. CES estimation utilizes a link relative estimation approach, where the
over-the-month change in weighted sample employment for a sample cell is
applied to the prior month’s estimate.
Non-CES area, referred to as non-covered areas (NCA), estimates are an optional
activity many states choose to generate in ACESweb each month, either for their
own analysis, publication, and/or as a LAUS input. The generation of NCA
estimates often requires considerable intervention due to small sample sizes and
bias among reporters. Traditional methods of generating the NCA estimates in
ACESweb often resulted in original figures that poorly reflected monthly
employment levels.
The Composite Quota Estimator (CQE) provides States with a more standard
method for producing monthly NCA estimates. This method incorporates
historical information and accounts for both QCEW seasonal monthly movement
and non-covered employment (NCE) movement. For LAUS purposes, the CQE
can be utilized as either an alternative for States to develop their own non-CES
Area (NCA) estimates or as an analytical tool to compare and track the NCA
estimates that States are currently producing. The use of the CQE does not
obligate States to replace their current method of creating NCA estimates. States
need to coordinate with their Regional Offices to have estimates for NCAs set up
in ACESweb.

Handbook line 1 input: Labor-Management Disputants (M02)
In addition to the non-agricultural wage and salary employment input (M01), data
for workers absent from their jobs due to labor-management disputes are needed.
Workers involved in these disputes can be either on strike (a work stoppage
initiated by the workers of an establishment) or lockout (a work stoppage initiated
by the management of an establishment). In both cases, the workers have jobs
from which they are temporarily absent and are therefore considered employed
under the CPS definition. (See Chapter 2.)
Counts of workers involved in labor disputes during the reference week are
entered into LAUSToo for each Handbook area using the variable ID M02.
Data on labor-management disputes can be obtained from the CES Strike Report
and the BLS Work Stoppages Program.

LAUS Program Manual 7-11

Handbook line 1 input: Dynamic Residency Ratios
The input data detailed in the previous section pertain to workers by
their place of employment. The CPS definition of employment pertains
to employed individuals by their place of residence. Because the goal of
LAUS is to parallel CPS concepts, the establishment data gathered and
entered into LAUSToo must undergo adjustment.
While there are several differences between establishment data and household
data (see Chapter 4), the largest source of difference at the M01 input area level is
the discrepancy between the location of establishments, where jobs are counted,
and the location of residencies, where employed individuals live. The Dynamic
Residency Ratios (DRRs) adjust the establishment-based inputs to a place-ofresidence, or household, basis.
Development of Residency-adjustment Methodology
Prior to 2005, residency adjustment of the establishment-based M01 and M02
inputs was accomplished by a single ratio for each LMA. The ratio was
calculated from 1990 Census data by dividing the number of residents employed
in non-agricultural wage and salary jobs within each LMA (obtained from Census
data) by the total number of wage and salary jobs in the LMA at the time of the
Census.
Beginning in 2005, the single-ratio approach was replaced with the introduction
of DRRs. The general concept behind the DRR methodology is that an LMA’s
resident employment is a function not only of the jobs available within the LMA
(the pre-2005 approach), but also of the jobs available in neighboring LMAs.
Commutation data from the 2000 Census long form were used to determine the
appropriate neighboring LMAs to include in the residency adjustment calculations
of each area. The largest commuter areas for each LMA were identified and, to
reduce the complexity of the calculations, the number of commuter areas was
capped at four for each LMA. A minimum of 100 commuters was set for a
commutation area to be included (for New England, the minimum was lowered to
50 commuters).
In 2015, several additional changes were introduced stemming from the
elimination of the Census long form and its replacement with the American
Community Survey (ACS).
When LAUS implemented DRRs, an important consideration was that it would be
at least a decade until the DRR commuting areas would be updated using worker
flows data from the next decennial census. For this reason, a relatively low
threshold for the inclusion of commuting areas was applied in order to capture as
LAUS Program Manual 7-12

much change as possible in area commuting patterns over the ten-year time frame.
With the replacement of the Census long form by the American Community
Survey (ACS), the time between releases of worker flows estimates was reduced
to every five years for the purpose of the OMB metro and micro area geography
updates. It was determined that 10% of total commuter employment was a
preferable threshold for including commutation areas in a residence area’s DRR
calculation. This new threshold achieved a desirable balance between operational
streamlining and capturing as much economic information as necessary.
The DRR inputs and calculations are detailed below.
Inputs for DRR calculations:
1. ACS worker flows data


County-to-county (or, in New England, MCD-to-MCD) commuter
flows from the ACS are aggregated to the M01 input area level.



The following are incorporated into the DRRs for each M01 input
area:
 Commuters residing and working within the same M01 input
area, regardless of level of commuters, and
 Up to the four largest commuter flows to neighboring M01
input areas, where the level of commuters is 10% or more of
total commutation.

2. ACS Non-agricultural Wage and Salary Employment (ACS NAWS)


Total ACS employment is obtained and the following are subtracted:
 Agricultural workers,
 Self-employed workers in own not incorporated business,
 Unpaid family workers, and
 Private household workers.

3. Establishment-based NAWS Employment base


M01 and M02 average for the same 5-year time period as the
corresponding ACS worker flows data set.

4. Annual Population Estimates (PEP population)


The Census Bureau’s Population Estimates Program (PEP) creates
annual total population estimates pertaining to July 1st of each year
following the decennial Census. These estimates are called postCensus estimates. For July 1st of years between completed Censuses,
PEP creates inter-Census estimates. PEP population for the same 5year time period as the corresponding ACS worker flows data set are
used along with ACS population to ensure proper population
controlling of ACS data.

5. ACS Population
LAUS Program Manual 7-13



Total population data from the ACS are typically but not always
aligned with the PEP population described above. Because the PEP
data are considered official, BLS uses them to adjust ACS data when
PEP and ACS population diverge.

Handbook line 1 input: Residency-based ACS Non-agricultural
Wage and Salary (NAWS) Employment (F01)
ACS NAWS data are used in the computation of DRRs, as noted above, and are also used
in a later disaggregation step when the geography of the M01 input encompasses two or
more Handbook areas. The disaggregation step apportions residency-adjusted NAWS
data at the M01 area level to the Handbook area level. The ACS NAWS data used for
this are adjusted on an annual basis using the latest available PEP data.

01

The following equation displays the DRR calculations for Residence Area1:

DRR 1:

⋯

DRR 2:

⋯

DRR n:

⋯

LAUS Program Manual 7-14

The following table displays an example DRR calculation based on the equations above:
(A)

(B)

(C)

(D)

(E)

(F)

Area

ACS
NAWS
Emp.

ACS
Commuters
from Area
of
Residence

Est.-based
NAWS
Emp.

5-year
average
PEP
population

ACS
population

235,584

305,400

590,249

590,116

1.0616

0.818913

43,370

127,450

590,249

590,116

1.0616

0.361252

Area of
Residence
Commuter
Area 1

296,071

(G)
Control
ratio
= [(B) for
Area of
Residence]
/ Σ (C) *
(E) / (F)

(H)

DRR
= (C) / (D)
* (G)

The following table displays an example Handbook line 1 calculation using the DRRs
from the table above where the geography of the M01 input is coterminous with a given
Handbook area:
(A)

Area

+
=

(B)

(C)

DRR

Establishmentbased NAWS
Employment
(M01)

Area of
Residence
Commuter
Area 1

(D)

(E)

Labor disputants
(M02)

Residencyadjusted line 1
component
= (B) * [(C) +
(D)]

0.818913

311,508

0

255,098

0.361252

121,105

200

43,822

Line 1 for Area of Residence

298,920

If the M01 input area above were to contain multiple Handbook areas, an additional
disaggregation step would be built into the line 1 calculation as follows:
(A)

Area

Area of
Residence
Commuter
+
Area 1
=

(B)

(C)

(D)

(E)

(F)

DRR

Establishment
-based
NAWS
Employment
(M01)

Labor
disputants
(M02)

ACS NAWS
(F01) for
Handbook
Area

ACS NAWS
(F01) for
M01 Area

(G)
Residencyadjusted line
1 component
= (B) * [(C)
+ (D)] * [(E)
/ (F)]

0.818913

311,508

0

45,012

308,102

37,268

0.361252

121,105

200

45,012

308,102

6,402

Line 1 for Area of Residence

43,670

LAUS Program Manual 7-15

All-other Employment (Handbook line 2)
All-other employment includes the following types of workers that are not
employed in agriculture:
(1) The self-employed, who work in their own not-incorporated business,
(2) Unpaid family members, who work for a business owned by a family member,
and
(3) Private household workers (or “domestic workers”).
These people are employed by the CPS definition and are the second largest
category of total employment behind non-agricultural wage and salary workers.
Two sources of all-other employment data exist--the American Community
Survey (ACS) and the CPS. The ACS provides more geographic detail while the
CPS is available on a monthly basis. Total all-other employment is calculated
using CPS estimates of all-other employment and ACS estimates of all-other
employment data.
All-other employment from the CPS are published on a monthly basis only for the
nation as a whole while unpublished data are available at the State level.
Research has shown that the State CPS data can be used as a control total at the
statewide level to develop ACS shares of all-other employment for the Handbook
areas.
Development of All-other Employment Methodology
The CPS and the ACS were identified as the most appropriate sources for the allother employment data that were previously obtained from the decennial Census
long-form survey. For LAUS purposes, both ACS and CPS offer differing
strengths and drawbacks. The goal is to utilize the strengths and to mitigate the
drawbacks of each data source. For instance, the CPS data are only available at
the State level, while the ACS data are available at the detailed geographic level
needed for LAUS substate Handbook method estimation. Also, the CPS data are
current and are available on a monthly basis, while the required ACS data are
available on a yearly basis in the form of 5-year estimates.
Since the monthly CPS data are only available at the State level and the ACS
provides more geographically detailed data, the ACS data are used to distribute
the CPS data to substate areas. To do this, the ACS all-other employment
estimate for a given area is divided by the sum of ACS all-other employment for
all areas within the State. The resulting ratio for a given area is referred to as the
“ACS share”.

LAUS Program Manual 7-16

The ACS share is expressed as:

Where:

area estimate of ACS all‐other employment
∑
sum of areas’ estimates of ACS all‐other
employment
The ACS shares are used to disaggregate CPS monthly statewide all-other
employment to the area level. The precedent for disaggregating statewide CPS
data comes from the Handbook methodology used to estimate new entrant and
reentrant unemployment. This method assigns a portion of the CPS statewide
new entrant and reentrant unemployment to individual areas based on a
population-specific ratio derived for the specific area.
CPS all-other employment estimates at the State level tend to be volatile monthto-month and are not suitable for direct use. To mitigate the volatility of the CPS
monthly statewide all-other employment estimates and obtain inputs more
suitable for handbook estimation, five years of CPS data for a given month are
used to develop weighted-average estimates. This allows the current month’s
CPS estimate to gain strength from prior year estimates while retaining the
seasonality of the reference month (The following table shows the weights used,
where “y” is the current year.). For consistency, the sum of 5-year State weighted
averages is controlled to the current monthly national CPS estimate of all-other
employment.
Year
y
y–1
y–2
y–3
y–4

Weight
0.40
0.25
0.20
0.10
0.05

Using a weighted average of statewide CPS all-other employment and the area
ACS share to generate the Handbook area all-other employment estimate is
expressed as follows:
All‐Other Emparea CPSwto * ACSshareo * CPSr
Where:
CPSwto Weighted average of the given month’s CPS all‐
other employment for the state
ACSshareo ACS share of all‐other employment for the
area
LAUS Program Manual 7-17

CPSr Ratio for controlling sum‐of‐State weighted
averages to national CPS all‐other employment

Agricultural Employment (Handbook line 3)
Unlike the non-agricultural Handbook employment estimates, which split
employment by class of worker—wage and salary (line 1) and “all-other” (line
2)—the agricultural Handbook employment estimate encompasses all classes of
worker—wage and salary, self-employed, and unpaid family—in a single
estimate.
The CPS and the ACS were identified as the most appropriate sources for the
agricultural employment data that were previously obtained from the decennial
Census long-form survey. As noted in the prior section, both ACS and CPS offer
differing strengths and drawbacks for LAUS purposes. The goal was to utilize the
strengths and to mitigate the drawbacks of each data source. For instance, the
CPS data are only available at the State level, while the ACS data are available at
the detailed geographic level needed for LAUS substate Handbook method
estimation. Also, the CPS data are current and are available on a monthly basis,
while the required ACS data are available on a yearly basis in the form of 5-year
estimates.
Since the monthly CPS data are only available at the State level and the ACS
provides more geographically detailed data, the ACS data area used to distribute
the CPS data to substate areas. To do this, the ACS agricultural employment
estimate for a given area is divided by the sum of ACS agricultural employment
for all areas within the State. The resulting ratio for a given area is referred to as
the “ACS share”. The ACS share is expressed as:

Where:

area estimate of ACS agricultural employment
∑
sum of areas’ estimates of ACS agricultural
employment
The ACS shares of agricultural employment are relatively stable from year to year
and are used to disaggregate CPS monthly statewide agricultural employment to
the area level. The precedent for using ACS data to disaggregate CPS agricultural
employment comes from the Handbook methodology used to estimate new
entrant and reentrant unemployment. This method assigns a portion of the CPS
statewide new entrant and reentrant unemployment to individual areas based on a
population-specific ratio derived for the specific area.
LAUS Program Manual 7-18

CPS agricultural employment estimates at the State level tend to be volatile
month-to-month and are not suitable for direct use. To mitigate the volatility of
the CPS monthly statewide agricultural employment estimates and obtain inputs
more suitable for handbook estimation, five years of CPS data for a given month
are used to develop weighted-average estimates. This allows the current month’s
CPS estimate to gain strength from prior year estimates while retaining the
seasonality of the reference month the following table shows the weights used,
where “y” is the current year.). For consistency, the sum of 5-year State weighted
averages is controlled to the currently monthly national CPS estimate of
agricultural employment.

Year
y
y–1
y–2
y–3
y–4

Weight
0.40
0.25
0.20
0.10
0.05

Using a weighted average of statewide CPS agricultural employment and the area
ACS share to generate the Handbook area agricultural employment estimate is
expressed as follows:
Agricultural Emparea CPSwta * ACSsharea * CPSr
Where:
CPSwta Weighted average of the given month’s CPS
agricultural employment for the state

ACSsharea ACS share of agricultural employment for
the area
CPSr Ratio for controlling sum‐of‐State weighted
averages to national CPS agricultural employment

LAUS Program Manual 7-19

Handbook Area Unemployment
Unemployment comprises all persons who do not have a job, have actively looked
for work in the prior four weeks, and are currently available for work. Persons
who were not working and were waiting to be recalled to a job from which they
had been temporarily laid off are also included as unemployed.
Receiving benefits from an Unemployment Insurance (UI) program has no
bearing on whether a person is classified as unemployed in the CPS survey;
however, statistics from the UI system are the only current measure of
unemployment available with a high degree of geographic detail. Because of this,
LAUS makes extensive use of UI records in its Handbook estimation procedures
for unemployment. Differences between the CPS concept of unemployment and
Handbook estimates of unemployment are resolved via additivity controls to
statewide estimates. (See Chapter 9.)
The Handbook method breaks unemployment into two main components—
experienced unemployed and entrant unemployment—that are each subdivided
into sub-components. Total Handbook unemployment (line 16) is experienced
unemployed (line 11) plus new entrant and re-entrant unemployment (lines 15 and
13, respectively).
Handbook
Line

Description
1. Experienced Unemployed (lines 8 + 9 + 10)


11
8

Continued Claimants (lines 5 + 6 + 7)
o State UI
o Unemployment Compensation for
Federal Employees (UCFE)
o Railroad Retirement Board (RRB)

5
6
7



Exhaustees

9



Non-covered Agricultural Unemployment

10

2. Entrant Unemployment


New Entrants

15



Re-entrants

13

Total Handbook Unemployment (lines 11 + 13 + 15)

16

LAUS Program Manual 7-20

Experienced Unemployment (Handbook line 11)
The largest component of unemployment comprises people that were employed in
the civilian labor force immediately before their current spell of unemployment.
These people are called the experienced unemployed.
The Handbook method estimates this component using three subcomponents:
(1) Continued Claimants (line 8)


Monthly counts of people receiving unemployment benefits
during the reference week

(2) Exhaustee Unemployment (line 9)


Estimates of people remaining unemployed after exhausting UI
benefits

(3) Non-covered Agricultural Unemployment (line 10)



Estimates of unemployed agricultural workers that are
ineligible for UI benefits
Atypical method only applicable for some States

Continued Claimants (Handbook line 8)
Actual counts of current UI claimants under State UI programs, the
Unemployment Compensation for Federal Employees (UCFE) program, and the
Railroad Retirement Board (RRB) program are included in the count of total
continued claimants. Continued claimants are defined as persons without
earnings certifying to a compensated or non-compensated week of unemployment
during the reference week.
Continued Claimants
Line

5

+

6

Input description
(LAUSToo variable ID)

Line Description

State UI
Continued Claims

UCFE Continued
Claims

Input
Source

Regular UI claims (M03)

State

Interstate UI claims (M04)

State

Commuter UI claims (M05)

State

Regular UCFE claims (M06)

State

Interstate UCFE claims (M07)

State

Commuter UCFE claims (M08)

State

RRB claims (M09)

BLS

+

7

RRB Continued
Claims

=

8

Total Continued Claims

LAUS Program Manual 7-21

Unemployed Exhaustees (Handbook line 9)
Final payments to beneficiaries of State UI and UCFE programs are tracked by
week and form the main input for exhaustee estimation. The exhaustee
component represents a significant part of overall unemployment estimates and is
a major contributor to inter-area variability in the estimates. Although States
know the number of individuals who receive final payments, they are unable to
track them after they leave the UI system. Each month, tabulations of weekly
counts of persons who have received final payments from the UI system are used
to estimate the number of people who do not immediately find a job or
discontinue their job search after exhausting benefits.
Estimates are made to reflect unemployment in the same CPS reference week as
continued claims, that is, the week including the 12th of each month. (In some
years, the November and\or December reference weeks are the week including the
5th of the month. LAUS technical memoranda inform States when this occurs.)
In addition, persons receiving final payments in previous weeks or months are
carried forward into successive periods in decreasing numbers by applying a CPSbased “survival” or continuation rate. This rate refers to individuals who are still
actively seeking and available for work. The estimate of current exhaustees for an
area is therefore “built up” over the period including the 19th of the previous
month through the week including the 12th of the current month, and includes an
estimate of the prior month’s unemployed exhaustees who remain unemployed in
the current month. The level of the pool of exhaustees can rise or fall depending
on the volume of final payments and the survival rate.
Inputs for Unemployed Exhaustees (line 9)
General Input Description

Weekly UI and UCFE
Final Payments

Quarterly Survival Rates

Specific Input description
(LAUSToo variable ID)

Input
Source

Regular (M10)

State

Interstate (M11)

State

Commuter (M12)

State

Rate group limits (S05 – S08)

BLS

Survival rates (S13 – S16)

BLS

Development of Exhaustee Methodology
Prior to 1987, a national average long-term survival rate was used to estimate the
number of unemployed persons that had exhausted UI benefits. The survival rate
was based on a formula developed by Hyman Kaitz that used national annual
average CPS duration data as the prime input. The underlying premise which
establishes the efficacy of the Kaitz method can be stated as follows: There is a
close and parallel relationship between the rate of unemployment and the duration
of unemployment spells (i.e., the survival rate). However, the application of a
single national annual average survival rate to all States and areas, regardless of
LAUS Program Manual 7-22

recent local unemployment rate conditions, does not fully conform to Kaitz’s
theoretical model. Therefore, beginning in 1987, a more flexible, timely, and
effective application of the basic Kaitz long-term survival rate methodology was
made operational.
That method established an unemployment rate-based survival rate that can
change at the area-level from month to month. Each quarter, the fifty States and
the District of Columbia are divided into four unemployment rate groups. Each
group represents a set of States within a given range of unemployment rates. In
addition, each group contains roughly twenty-five percent of the nationally
weighted unemployment. Using the Kaitz formula and quarterly average CPS
State duration data, a survival rate is developed for each of the four
unemployment rate ranges. Thus, on a monthly basis, areas will be assigned a
survival rate that most closely relates to recent local unemployment rate
conditions. In this manner, high unemployment rate areas select a higher survival
rate and have higher exhaustee levels and Handbook unemployment estimates
than low unemployment rate areas.
Research established a lagged correlation of two quarters between the
unemployment rate and the survival rate. Adding an operational lag of one
quarter results in the use of a given survival rate based on the area’s
unemployment rate nine months prior.
In implementing this procedure, the following occurs:
1. Every January, April, July, and October, four survival rates are issued
based on CPS data for the most recent quarter (4th, 1st, 2nd, and 3rd).
2. Each month during a given quarter, areas are assigned a survival rate for
the quarter of receipt (1st, 2nd, 3rd, and 4th).
3. The selection of the rate is based on the area’s total unemployment rate
nine months prior to the estimate month. This lag represents the two
quarter lagged relationship between the unemployment rate and the
survival rate plus a one-quarter operational lag.
Survival Rate Calculation
Survival rates are calculated using the 15-26 weeks unemployed and the 27 weeks
and over unemployed duration data. The two classes of duration data are assumed
to have an exponential distribution and the probability of falling into each of the
classes is calculated based on that assumption by taking the integral of the
exponential distribution function between the boundaries of each class (15-26
weeks or 27 weeks and over). The equation for the probability of failing into
either class is known based on this calculation and the distribution of the two
duration classes is known from the CPS duration data. All that is needed is to find
the death rate (the likelihood that someone leaves the exhaustee pool each week)
which maximizes the likelihood of having the distributions given in that quarter’s
LAUS Program Manual 7-23

CPS duration data for the two classes, given the probability of falling into each
class. The first derivative of the likelihood function is set to zero and solved for
the death rate in order to find the local maximum. This maximum is equivalent to
the death rate that maximizes the likelihood of having the distribution of duration
classes given by the CPS data given the probabilities of a person falling in each
class. This death rate is then used to calculate the long-term survival rate, which is
used to calculate the exhaustee data. (See Apendix 7-1 for more details.)
Calculation of Exhaustees from Weekly Final Payments
The following two tables illustrate the steps involved in calculating exhaustees.
The example in the first table pertains to the two-month estimation period for
March (revised prior month) and April (preliminary current month) of 2009.
Each column of the worksheet is described in detail in the second table.
(A)

(B)

(C)

(D)

(E)

Reference
Month

Week
Ending
Date

Total
Final
Payments

Survival
Rate

Exhaustee
Estimate1

February

2/14/2009

31

2/21/2009

30

0.959

380 = ( 31 + 365 ) * 0.959

2/28/2009

29

0.959

393 = ( 30 + 380 ) * 0.959

3/7/2009

16

0.959

405 = ( 29 + 393 ) * 0.959

3/14/2009

17

0.959

403 = ( 16 + 405 ) * 0.959

3/21/2009

16

0.955

401 = ( 17 + 403 ) * 0.955

3/28/2009

17

0.955

399 = ( 16 + 401 ) * 0.955

4/4/2009

32

0.955

397 = ( 17 + 399 ) * 0.955

4/11/2009

27

0.955

410 = ( 32 + 397 ) * 0.955

0.955

417 = ( 27 + 410 ) * 0.955

March
(rev.)

April
(prelim.)

4/18/2009

(F)
Exhaustee Calculation
= [(C) + (E)]prior week x
(D)current week

365 = Starting Pool

1

The exhaustee estimate for the reference week is rounded to the whole integer and becomes the monthly
Handbook Line 9 value. For the weeks between reference weeks, the exhaustee estimates are unrounded.
Rounded values are displayed here for ease of reference.
LAUS Program Manual 7-24

Column
(A)

(B)

(C)

(D)

Description
Reference Month


The months for which estimates are being generated.



In the example, the estimation period pertains to March (revised) and April
(preliminary) of 2009. Estimates for these periods were created during the month
of May 2009. The last week of February is included to display the prior month’s
final payments and exhaustee pool.

Week Ending Date


The last day of each calendar week displayed by reference month. The last week
of each reference month is the CPS reference week (usually the week including
the 12th day of the month).



Each reference month starts in the week following the prior month’s reference
week and extends to the current reference week.

Final Payments


The count of people receiving their last unemployment benefit payment during
the week indicated.



Final payments made from State UI programs and from the UCFE program are
included. RRB final payments are not tracked.



Note that final payments made during the reference week are also counted as
continued claims.

Survival Rate


(E)

(F)

The rate at which exhaustees and final payment recipients from the prior week
remain unemployed in the current week.

Exhaustee Estimate


A weekly estimate of the number of unemployed people who have exhausted UI
benefits.



The exhaustee estimate for the reference week is the monthly Handbook line 9
estimate.

Exhaustee Calculation


The survival rate for the current week is applied to the prior week’s sum of final
payments and exhaustee pool to yield the exhaustee pool for the current week.
∗

Where “n” is the current week and “n-1” is the prior week.

LAUS Program Manual 7-25

Non-covered Agricultural Unemployment (Handbook line 10)
Generally, this component is a small part of unemployment,
but it is very important for some areas with large and highly
seasonal agricultural sectors. For the 17 States that estimate
non-covered agricultural unemployment, this component
accounted for less than one percent of total Handbook
unemployment; however, for individual areas within those
17 States, the component accounted for up to a quarter of the
unemployed.
Direct estimation of agricultural unemployment may be used in States with at
least one Handbook area where agricultural employment is 25 percent or more of
total employment. States that qualify must obtain approval from BLS to estimate
agricultural unemployment directly. In such cases, this direct estimation must be
used in all Handbook areas of the State. Other States may request approval for
atypical treatment of agricultural unemployment for a specific Handbook area if it
can be demonstrated that the lack of such an estimate has a deleterious effect on
estimates for that area.
The following formula details the calculation of non-covered agricultural
unemployment. The table below the formula provides details.
If A01 > L03, then L10 = 0, otherwise
L08 L09
L03 A01 ∗
L04 L08 L09
L10
L08 L09
1
L04 L08 L09
Where:
Identifier
A01

Description
Covered Agricultural Employment


Obtained from monthly QCEW data. When the current
month’s QCEW data are unavailable, data from the same
month one year ago are used. Revisions will incorporate
the current month’s data when they are available.

L03

Handbook Agricultural Employment

L04

Total Handbook Employment

L08

Total Continued Claims

L09

Exhaustee Unemployment

L10

Non-covered Agricultural Unemployment
LAUS Program Manual 7-26

Entrant Unemployment (Handbook lines 13 and 15)
For many unemployed individuals, their current spell of unemployment was not
immediately preceded by employment. These individuals entered the labor
market from outside the labor force after having completed military service,
family responsibilities, education, or other activities outside the civilian labor
force. These individuals are known as unemployed entrants.
Unemployed entrants are further divided into two groups:
(1) Unemployed New Entrants


Individuals who enter the labor market for the first time and do not
find jobs.

(2) Unemployed Re-entrants


Individuals who enter the labor market after a period of retirement
from the labor force and are unable to find jobs.

Estimates of new entrants and re-entrants are created for each State, and the
statewide estimates are then allocated to the Handbook areas within the State
using annual population data. The table below lists the inputs necessary for
entrant estimation in LAUSToo.
Inputs for Unemployed Entrants
Input/ Handbook Line description
(LAUSToo variable ID)

Line

13

15

Input
Source

Statewide Unemployed Re-Entrants (SRE)

BLS

x

Re-entrants allocation ratio

BLS

=

Unemployed Re-entrants ( = SRE * Re-entrants
ratio)
Statewide Unemployed New Entrants (SNE)

BLS

x

New Entrants allocation ratio

BLS

=

Unemployed New Entrants ( = SNE * New
Entrants ratio)

Statewide New Entrant and Re-entrant Unemployment (SNE and SRE)
Statewide new entrant and re-entrant estimates are available from the CPS;
however, the data are volatile and are not suitable for direct use. To reduce the
volatility of the CPS estimates and obtain inputs more suitable for Handbook
estimation, five years of CPS data for a given month are used to develop
weighted-average estimates. The following table shows the weights used, where
“y” is the current year.
LAUS Program Manual 7-27

Year

Weight

y
y–1
y–2
y–3
y–4

0.40
0.25
0.20
0.10
0.05

Once the weighted estimates are calculated, the resulting data for the fifty states
and the District of Columbia are controlled to monthly national CPS estimates of
new entrant and re-entrant unemployment. This controlling step, which was
added to the methodology in 2010, ensures that use of data from earlier years does
not bias the overall level of Handbook entrant unemployment upwards or
downwards during times of generally falling or rising unemployment,
respectively.
Allocation of Statewide Entrants to Handbook Areas
Handbook calculations in LAUSToo distribute the statewide new entrant and reentrant estimates to the Handbook area level using each area’s share of statewide
age-group population data. BLS obtains the population data from the U.S. Census
Bureau each year. New entrants are distributed using Handbook area shares of
the population aged 16 to 19 years, while shares of the population aged 20 or
more years are used to allocate re-entrants.

LAUS Program Manual 7-28

Appendix 7-1
Survival Rate Formula
1. An exponential distribution was fitted to the distribution classes of “15 to 26 weeks” and
“27 weeks and over”. (The actuals bounds will be listed as 14.5 to 26.5 weeks and 26.5
weeks and over.) The exponential density function is:
λ

∞ 

           0

where:
λ = the single parameter to be estimated by the method of maximum
likelihood (the survival rate)
t = duration in weeks 

2. The probability of falling in each of the two classes is determined by integrating over the
density:
where:
λ = the single parameter to be estimated by the method of maximum
likelihood (the death rate)
t = duration in weeks 

λ

where:
TU = upper class boundary
TL = lower class boundary
Prob (falling in 14.5-26.5) =
Prob (falling in 26.5+) =

.

.

.

These locations are shifted by 13.5 weeks so that:
Prob (falling in 14.5-26.5) =
Prob (falling in 26.5+) =

Doing so indicates that the chances of surviving or not surviving in the
exhaustee pool do not begin until week 13.5 for the purposed of
determining the exhaustee rate. A result of this is that the function is now
binomial, in that a person will only fall in one of the two duration classes.
3. The likelihood function (L), as used here, is defined as the joint probability (or product)
of the probabilities of a person falling in either of the two duration classes. Maximizing L
with respect to λ yields the maximum likelihood estimate of this parameter. The
likelihood function appropriate here is the binomial:
LAUS Program Manual 7-29

!
!

!

where:
k1 = number of unemployed in 14.5-26.5 week class
k2 = number of unemployed in 26.5+ week class
K = total number unemployed in both classes (k1 + k2)
This equation is derived from the following:

1

is the likelihood function for a binomial process.
is the combination of choosing ‘x’ from a set of ‘n’.

!
!
!

In this case we have:

!

!

where (K-k1) is equivalent to k2

or the probability of falling into the 14.5-26.5 week class
1

or the probability of falling into the 26.5+ week class

4. The natural logarithm (ln) of the likelihood function (L*) is obtained because it is easier
to work with and because the maximum of the likelihood function will occur at the same
value of λ as will the maximum of the natural logarithm of the likelihood function
∗

ln

!

ln

!

ln

!

ln

ln

5. The derivative of L* is obtained with respect to λ and set equal to zero to obtain the
maximum likelihood estimate of λ. An equation will reach a local maximum or minimum
where its first derivative is equal to zero. In this case, we are seeking to take the derivate
of the above equation with respect to λ in order to find the value of λ for which the
equation equals zero. This will be the maximum and will be the value of λ which
maximizes the likelihood of the given distribution of the two duration classes.
∗

13

13

λ

0

This can be written as: 

13

13

13

1

LAUS Program Manual 7-30

Solving for λ yields:
1
ln 13
λ
12

1

ln 13

13

6. The long-term survival rate (Pi) is a constant. The formula used is: 

for all i
where λ has been estimated in Step 5 and i is the weeks duration
for the 1-13 and 13+ weeks duration scale.
Thus the survival rate formula is as follows:

LAUS Program Manual 7-31

8 Geography

Introduction
he Local Area Unemployment Statistics (LAUS) program is
responsible for estimation of unemployment rates for areas below
the national level. This chapter will provide a comprehensive
review of LAUS geography, beginning with the largest areas and
proceeding through the area types in descending order by size.

T

For each type of geography, the source of the delineation, the role of the
geography in LAUS estimation, and the estimation methodology will be
noted. Detailed information regarding each estimation methodology can
be found in the appropriate chapters.

LAUS Program Manual 8-1

Census Regions and Divisions
The U.S. Census Bureau has designated four census regions and nine
divisions for which LAUS creates labor force estimates. The four census
regions each comprise two or more census divisions. The nine census
divisions each comprise three or more states. The table below summarizes
the geographic composition of these areas.
LAUS creates estimates for census divisions using statistical models that
incorporate data from the Current Population Survey (CPS). These
models are benchmarked, or forced to sum to, the national estimates on a
monthly basis. The benchmarked census division estimates are then used
as benchmarks for their component states. Estimates for census regions
are developed by summing the model-based data of their component
census divisions.
Note that Puerto Rico, which is within the scope of the LAUS program,
is not part of any census region or division.
Census Region
Northeast

Midwest

South

West

Census Division

States

New England

Connecticut, Maine, Massachusetts, New
Hampshire, Rhode Island, and Vermont

Middle Atlantic

New Jersey, New York, and Pennsylvania

East North
Central

Illinois, Indiana, Michigan, Ohio, and
Wisconsin

West North
Central

Iowa, Kansas, Minnesota, Missouri, Nebraska,
North Dakota, and South Dakota,

South Atlantic

Delaware, District of Columbia, Florida,
Georgia, Maryland, North Carolina, South
Carolina, Virginia, and West Virginia

East South
Central

Alabama, Kentucky, Mississippi, and
Tennessee

West South
Central

Arkansas, Louisiana, Oklahoma, and Texas

Mountain

Arizona, Colorado, Idaho, Montana, Nevada,
New Mexico, Utah, and Wyoming

Pacific

Alaska, California, Hawaii, Oregon, and
Washington

States
LAUS publishes labor force estimates for the 50 states, the District of
Columbia, and the Commonwealth of Puerto Rico. (While the District and
Puerto Rico are not states, the LAUS program generally treats them as
such for administrative purposes.)
Estimates for 48 states and the District of Columbia are developed using
statistical models. In addition to using CPS data as the primary input, these
models also incorporate Current Employment Statistics (CES) total
LAUS Program Manual 8-2

nonfarm employment data for the employment model and unemployment
insurance (UI) claims data for the unemployment model. State model
estimates are controlled, or forced to sum to, their respective census
division estimates.
The two remaining states—California and New York—are treated
differently. Due to their large sizes, both the modeled substate areas and
the balances of state are treated as states for modeling purposes, summed
to create labor force estimates for California and New York, and
controlled directly to their respective census division estimates.
Estimates for Puerto Rico are not modeled, but directly derived from a
separate survey similar to the CPS that is administered by the Puerto Rico
Department of Labor.
Balances of State
Balance of state is the portion that remains after removing the modeled
substate area within a state. There are seven balances of state that exist to
facilitate model-based estimation for areas within their respective states.
LAUS creates estimates for balances of states and their respective substate
modeled areas using statistical models. Each month the balance of state
estimates, along with the respective modeled substate area estimates, are
controlled to the state model estimates.
As previously noted, the balance of state and substate modeled area
estimates for California and New York are treated as states and are
controlled directly to their respective census division estimates.

State

Balance of State

Substate Modeled Area

California

Balance of California

Los Angeles-Long Beach-Glendale,
CA Metropolitan Division

Florida

Balance of Florida

Miami-Miami Beach-Kendall, FL
Metropolitan Division

Illinois

Balance of Illinois

Chicago-Naperville-Arlington Heights,
IL Metropolitan Division

Michigan

Balance of Michigan

Detroit-Warren-Dearborn, MI
Metropolitan Statistical Area

New York

Balance of New York

New York city, NY

Ohio

Balance of Ohio

Cleveland-Elyria, OH Metropolitan
Statistical Area

Washington

Balance of Washington

Seattle-Bellevue-Everett, WA
Metropolitan Division

Labor Market Areas
In the late 1940s, when subnational labor force estimation was first
LAUS Program Manual 8-3

attempted, employment and unemployment estimates were developed for
large labor market areas (LMAs) as well as for states, underscoring the
importance of substate labor market information. Subsequently, all LMAs
were identified in order to standardize and promote comparability for the
collection and use of labor force information in administering various
government programs. In the LAUS program, substate estimates of
employment and unemployment are prepared for all LMAs in the Nation.
Generally, an LMA is defined as an economically integrated geographic
area within which individuals can reside and find employment within a
reasonable distance or can readily change employment without changing
their place of residence. LMAs are either metropolitan areas, micropolitan
areas, or small LMAs, and they exhaust the geography of the U.S. and
Puerto Rico.
The Office of Management and Budget (OMB) is responsible for
delineating metropolitan and micropolitan areas for use by federal
statistical agencies in collecting, tabulating, and publishing federal
statistics. As of 2015, there are:


366 Metropolitan Statistical Areas, 10 of which contain 28
Metropolitan Divisions, and 524 Micropolitan Statistical Areas in
the non-New England states.



21 Metropolitan New England City and Town Areas (NECTAs),
1 of which contains 10 NECTA Divisions, and 17 Micropolitan
NECTAs in the 6 New England states.



7 Metropolitan Statistical Areas and 5 Micropolitan Statistical
Areas in Puerto Rico.

The LAUS program is responsible for delineating small LMAs. LMAs are
delineated in terms of counties or county equivalents in all areas except
New England, where Minor Civil Divisions (MCDs) are used.
LMAs are delineated on the basis of population, urbanization, and
commutation data. Since population and urban area data are inappropriate
for delineating the generally less populous small LMAs, commutation
data are used to determine which counties are deemed single-county
LMAs and which are combined into multi-county LMAs. Regardless of
population size, commuting flows are an indication of the degree of
integration of labor markets among counties.

LAUS Program Manual 8-4

Federal Statistical Areas
Metropolitan Statistical Areas and Micropolitan
Statistical Areas collectively are called Core Based
Statistical Areas, or CBSAs. The Metropolitan and
Micropolitan Statistical Area Standards do not equate
to an urban-rural classification; most counties
included in Metropolitan and Micropolitan Statistical Areas and many
outside-CBSA counties contain both urban and rural territory and
populations.
Core
A core is a densely settled concentration of population, comprising either
an urbanized area of at least 50,000 population or an urban cluster of
10,000 to 49,999 population determined by the Census Bureau, around
which a CBSA is defined. (See www.census.gov/geo/reference/ua/urbanrural-2010.html for information from the Census Bureau on urban and
rural classification.)
Core Based Statistical Area (CBSA)
A CBSA is a statistical geographic entity consisting of the central county
or counties associated with at least one urban core of at least 10,000
population, plus adjacent, outlying counties having a high degree of social
and economic integration with the core as measured through commuting
ties. Metropolitan and Micropolitan Statistical Areas are the two
categories of CBSA. A Metropolitan Statistical Area is based on an
urbanized area core, while a Micropolitan Statistical Area is based on an
urban cluster core.
New England City and Town Area (NECTA)
A NECTA is a statistical geographic entity conceptually similar to a
CBSA, except that Minor Civil Divisions, or MCDs, are used as its
building blocks instead of counties. NECTAs are an alternative,
equivalent set of areas delineated by OMB for the six New England states,
in recognition of the primacy of cities and towns in local area governance
within that census division. There are Metropolitan and Micropolitan
NECTAs, depending on the population of the urban core.
Metropolitan Division and NECTA Division
A Metropolitan Division is a county or group of counties within a
Metropolitan Statistical Area that contains a core population of at least 2.5
million. A Metropolitan Division consists of at least one main county or
at least two secondary counties that represent an employment center, plus
adjacent counties associated with this employment center through
commuting ties. Metropolitan Divisions are subdivisions of very large
Metropolitan Statistical Areas that often function as distinct social,
LAUS Program Manual 8-5

economic, and cultural areas with the larger region, and they retain their
separate statistical identities. For a Metropolitan NECTA containing a
core population of at least 2.5 million, NECTA Divisions are delineated
where separate employment centers can be identified using MCDs as
building blocks.
Combined Statistical Areas and Combined NECTAs
A Combined Statistical Area consists of two or more adjacent
Metropolitan Statistical Area(s) and/or Micropolitan Statistical Area(s)
linked through commuting ties. Areas are combined based on an
employment interchange rate of at least 15 percent (either inflows or
outflows). A Combined NECTA consists of two or more adjacent
Metropolitan NECTA(s) and/or Micropolitan NECTA(s) similarly linked
through commuting ties.
The standards underlying the 2010 Census-based federal statistical area
delineations are available through the OMB website at
www.whitehouse.gov/sites/default/files/omb/assets/fedreg_2010/062820
10_metro_standards-Complete.pdf. Links to OMB update bulletins listing
statistical area delineations based on these standards are available at
www.census.gov/population/metro/data/omb.html.

LAUS Program Manual 8-6

Small Labor Market Areas
While CBSAs and NECTAs are delineated by the Office of Management
and Budget (OMB), LAUS delineates small LMAs over territory outside
of CBSAs (outside of NECTAs in the New England states). Similar to the
federal statistical areas, multi-entity small LMAs are created based on
commutation data from the American Community Survey (ACS) 5-year
average dataset. However, unlike the federal statistical areas, no
population criteria are applied in delineating small LMAs.
(1)

Worker flows are examined, and counties are combined into one
small LMA if either of the following conditions is met:
(a) At least 25.0 percent of the employed residents of one county
commute to work in another county.
(b) At least 25.0 percent of the employment (persons working) in
one county is accounted for by workers commuting from
another county.

(2)

Small LMAs, as is the case with metropolitan and micropolitan
areas, are required to be contiguous. Counties are first combined
based on the commutation criteria, and then potential multi-county
small LMAs are checked for contiguity. Noncontiguous portions
of potential small LMAs are considered separately. If the
noncontiguous area contains more than one county, it is
reevaluated using criteria (1)(a) and (1)(b) above. If the
noncontiguous area consists of a single county, it is delineated as
a separate small LMA.

(3)

Subsequent to the verification of contiguity described in (2) above,
commuting flows between adjacent small LMAs are evaluated.
Those areas for which the measures and thresholds specified in
(1a) and (1b) above are met merge to form one small LMA.

(4)

For the New England MCD-based small LMAs, due to the large
number of small MCDs, residual MCDs are added to contiguous
small LMAs based on commuting flows and/or other economic
ties. That is, if, after applying the commutation criteria, an MCD
is identified as an individual small LMA, the MCD is added to a
contiguous small LMA, especially if the MCD is very small.

A complete listing of the metropolitan, micropolitan, and small LMAs in
use by the LAUS program is available through the BLS website at
www.bls.gov/lau/lausmsa.htm.

LAUS Program Manual 8-7

Methodology for Labor Market Area Data
More than half of all non-New England counties and county equivalents
are coextensive with federal statistical areas or small LMAs. Estimates for
single-county Metropolitan Statistical Areas, Metropolitan Divisions,
Micropolitan Statistical Areas, and small LMAs are produced using the
Handbook method by virtue of their geographic equivalence to counties.
For multi-county areas, estimates are aggregated from the Handbookbased estimates of the component counties. Estimates for large
Metropolitan Statistical Areas containing Metropolitan Divisions
similarly are aggregated from their Metropolitan Divisions, while
estimates for Combined Statistical Areas are aggregated from their
Metropolitan and/or Micropolitan Statistical Area components.
In the New England states, the Handbook method is used at the MCD
level. No New England MCDs are coextensive with federal statistical
areas, while only two MCDs are coextensive with small LMAs. Hence,
estimates for all NECTA geography and virtually all small LMA
geography in New England are aggregated from the Handbook-based
estimates of their MCD components.
The largest component of household employment—nonagricultural wage
and salary, or NAWS—is often estimated at the labor market area level
and then distributed to the component counties or MCDs of multi-entity
labor market areas using American Community Survey data. Thus, labor
market area geography has an important role in the Handbook method.
Some states have elected to estimate NAWS at the county level for their
multi-county Micropolitan Statistical Areas and small LMAs.
Expanded Estimating Areas are unique to the New England states and
were created to facilitate NAWS estimation for small, isolated MCDs.
They comprise a Metropolitan or Micropolitan NECTA and at least one
such MCD deemed too small for effective NAWS estimation by itself.

LAUS Program Manual 8-8

Counties
LAUS creates estimates using the Handbook method for non-New
England counties and county equivalents in the U.S. and Puerto Rico.
(County equivalents include Boroughs and Census Areas in Alaska;
Parishes in Louisiana; Independent Cities in Missouri, Nevada, and
Virginia; Municipios in Puerto Rico, and various county/city areas in
other states). Estimates for counties in New England are aggregated from
the Handbook-based estimates of their component MCDs.
New England Minor Civil Divisions (MCDs)
In much of New England, counties are statistical entities only, as defined
by the Census Bureau. Governmental functions typically associated with
the county level elsewhere in the U.S., rather, tend to be carried out by
cities and towns (i.e., at the MCD level) in New England. As OMB
recognizes the importance of MCDs in its delineation of NECTAs as an
equivalent alternative to the county-based Metropolitan and Micropolitan
Statistical Areas, the LAUS program acknowledges the primacy of MCDs
in New England through its use of exhaustive Handbook estimation at the
MCD level there.
Incorporated Places and Minor Civil Divisions Outside of New
England
The LAUS program produces estimates for incorporated places with
populations of 25,000 or more, plus MCDs with populations of 25,000 or
more in the states of Michigan, New Jersey, New York, and Pennsylvania.
Disaggregation techniques are used to create estimates for these areas.
Each year, typically by the end of May, the Census Bureau issues
population estimates for sub-county areas with a reference point of July 1
of the prior year. (See www.census.gov/popest/index.html on the Census
Bureau website.) These data are reviewed to determine sub-county areas
newly meeting the 25,000-threshold for inclusion in the LAUS program.
The list of new areas is provided to the affected states for addition during
the next annual processing cycle. Sub-county areas falling below the
25,000-threshold generally are not considered for removal from the
LAUS program except during decennial redesigns.
Incorporated Place Parts
There are incorporated places for which LAUS (1) creates estimates and
(2) recognizes territory in more than one county. For any incorporated
place estimated by LAUS, data for the county-specific parts are produced
through disaggregation if the incorporated place has more than one county
part with a nonzero labor force level in the base American Community
LAUS Program Manual 8-9

Survey 5-year dataset. These parts are then summed to the whole
incorporated place. In a small number of cases, LAUS incorporated places
are split across three or more counties where some parts are recognized
but some part or parts are not, based on the American Community Survey
dataset in use.
State-Specific Areas
States are given the option at the end of each year to inform BLS if they
wish to add state-specific areas to the production database or modify
existing state-specific areas. These are typically sub-county areas below
the LAUS population threshold of 25,000—for which estimates generally
are developed through disaggregation techniques—or workforce regions,
for which estimates generally are developed through aggregation. BLS
does not publish estimates for state-specific areas, and each state assumes
this responsibility for its state-specific areas.
Areas of Substantial Unemployment
Areas of Substantial Unemployment, or ASUs, are (1) contiguous
geographic areas (2) with populations of at least 10,000 and (3)
unemployment rates of 6.5 percent or more. They are used by the
Employment and Training Administration (ETA) to allocate funds to
states under a provision of the Workforce Innovation and Opportunity
Act. Under the Cooperative Agreement, states are required to submit
ASUs to BLS through the ASU module of the production system each
year, typically by mid-October, which BLS then validates on behalf of
ETA. A state maximizes its allocation by configuring its ASUs to include
as a high a share of its total unemployed as possible. Historically, ETA
has permitted states to use census tracts to develop their ASUs, although
LAUS areas can be used as well. (The granularity that tract-level data
provides can be particularly useful when unemployment rates for LAUS
areas are generally low.) The ASU module of the production system
includes data for all census tracts for the sole purpose of facilitating ASU
development. These data reflect a weak disaggregation technique from the
county level (including in the New England states) based on volatile
American Community Survey 5-year estimates. They are not of BLS
publication quality and are not recommended for any other use.
Bureau of Labor Statistics Regions
The Bureau of Labor Statistics (BLS) has subdivided the Nation into
regions for administrative purposes. LAUS does not create estimates for
BLS regions.

LAUS Program Manual 8-10

BLS Region

States

Boston / New York

Connecticut, New York, Maine, Massachusetts, New
Hampshire, Rhode Island, Vermont, and Puerto Rico
Delaware, District of Columbia, Maryland, New Jersey,
Pennsylvania, Virginia, and West Virginia
Alabama, Florida, Georgia, Kentucky, Mississippi, North
Carolina, South Carolina, and Tennessee
Illinois, Indiana, Iowa, Michigan, Minnesota, Nebraska, North
Dakota, Ohio, South Dakota, and Wisconsin
Arkansas, Colorado, Kansas, Louisiana, Missouri, Montana,
New Mexico, Oklahoma, Texas, Utah, and Wyoming
Alaska, Arizona, California, Hawaii, Idaho, Nevada, Oregon,
and Washington

Philadelphia
Atlanta
Chicago
Dallas / Kansas City
San Francisco

LAUS Program Manual 8-11

9 LAUS Estimation: Additivity
Introduction
inking substate labor force estimates to the CPS concepts begins
with a set of Handbook employment and unemployment
estimates. These Handbook estimates are prepared for all
counties outside of New England and all Minor Civil Divisions
(MCDs) in New England. Because of nonlinearity in the Handbook, the
county or MCD employment and unemployment estimates will not
necessarily equal the statewide totals. Hence, an additivity adjustment
must be performed. This process introduces conformity between the
Handbook and the statewide estimates by making the sum of the
exhaustive county or MCD estimates additive to statewide level estimates.
The LAUS program uses a simple linear additivity adjustment method,
referred to as the Handbook-Share technique, to adjust county or MCD
estimates to the statewide estimates. This method consists of distributing
the statewide estimates, based on the percentage share of each Handbook
area estimate, over the total of the Handbook estimates. This method is
applied to all areas for which an independent Handbook estimate is
prepared. The adjustments for additivity are performed on a current basis,
and whenever the statewide estimates are revised.
After the Handbook estimates have been adjusted for additivity to the
statewide estimates, the estimates are referred to as “LAUS” estimates
rather than Handbook estimates.

LAUS Program Manual 9-1

When estimates for areas below the county are needed, the LAUS
estimates are then disaggregated into sub-county areas, such as cities and
towns. Two methods for disaggregation exist based on the availability of
UI claims and ACS data for apportioning county estimates to smaller
areas. See Chapter 10 for a complete description of the disaggregation
process.

LAUS Program Manual 9-2

Adjustment to Independent Statewide Estimates—
The Handbook Share Method
The process of reconciling, or linking, county or MCD labor force
estimates to Statewide (model-based) estimates begins with a set of
geographically exhaustive Handbook employment and unemployment
estimates. A simultaneous adjustment for additivity of all county or MCD
estimates to the statewide estimates is performed using the percentage
distribution of the substate Handbook estimates, also known as the
Handbook-Share method. The Handbook-Share method of apportioning
the State estimates of unemployment and employment to areas assumes a
proportional distribution throughout the State of the difference between
the sum of substate Handbook estimates and the independent State
estimates. This adjustment is performed for both preliminary and revised
estimates.
The following worksheet illustrates simultaneous additivity and
adjustments to counties and MCDs using the Handbook-Share method.

Simultaneous Additivity of County or MCD Estimates Using the
Handbook-share Method

Area

Unemployment

Employment

Percent of
Summed
Handbook

Percent of
Summed
Handbook

Handbook

State

Statewide*

Handbook

49,300

Statewide*

562,800

County/MCD1

18,500

0.394456

19,447

190,600

0.3481279

195,926

County/MCD 2

9,300

0.198294

9,776

107,100

0.1956164

110,093

County/MCD 3

8,700

0.185501

9,145

103,400

0.1888585

106,290

County/MCD 4

2,300

0.049041

2,418

36,800

0.0672146

37,828

County/MCD 5

1,900

0.045120

1,997

25,900

0.0493059

26,624

County/MCD 6

6,200

0.132196

6,517

83,700

0.1528767

86,039

Sum of substate
Areas

46,900

1.000000

49,300

547,500

1.000000

562,800

*For the State, enter the model-based estimate. The substate data are the product of the
area’s Percent of Summed Handbook and the statewide estimate for unemployment or
employment.

LAUS Program Manual 9-3

10

LAUS Estimation:
Disaggregation

Introduction
isaggregation techniques are used to obtain current
estimates of employment and unemployment for
cities and parts of cities where the city lies in more
than one county. Disaggregation involves prorating
employment and unemployment for one or more counties to
the disaggregated area. Since these areas are within counties, independent employment
and unemployment estimates cannot be developed, as basic data are not always available
and current LAUS estimating procedures are not applicable.

D

Disaggregation methods are used to develop estimates for cities with a population of
25,000 or more.
Two methods of disaggregation are appropriate for LAUS use.
1.) ACS-share employment disaggregation uses ACS 5-year employment estimates indexed
to the July-1 post-censal population estimates from the most recent vintage available during
the previous annual processing cycle.
2.) Claims-based unemployment disaggregation uses current UI claims data by city of
residence, 2010 Census population data by age group, and the most recent Census Bureau
population estimates.

The starting point for disaggregation is the estimate of employment and unemployment
prepared for the Handbook Area in accordance with Handbook instructions outlined in
Chapter 7 and the directions on adjustment for additivity to statewide totals in Chapter 9.

September 2015

LAUS Program Manual 10-1

ACS-Share Disaggregation
Since current employment and unemployment estimates at the city level are required to
implement numerous Federal economic assistance and employment and training
programs, methods of disaggregation which reflect current economic conditions in these
cities are necessary. Apart from the Census Bureau’s American Community Survey
(ACS), there are very few economic and demographic data series for small areas. Two
exceptions are the monthly UI claims series and the population estimates prepared by the
Census Bureau.
Specifically, the current data used in LAUS disaggregation are UI continued claims by
place of residence for the week including the 12th of the month and total population
estimates prepared annually pertaining to July 1 of the given year. The procedures which
incorporate the use of these data are known as ACS-share and claims-based
disaggregation.
ACS-Share Employment Disaggregation for Cities

American Community Survey employment estimates are available from the 5-year
datasets published annually by the Census Bureau for a comprehensive set of areas,
including all disaggregated cities and city parts. The most recent 5-year ACS
employment estimate for the city or city part is divided by the ACS employment estimate
for the corresponding county. Research showed that, while ACS uses Census Population
Estimates Program (PEP) population estimates as controls, ACS population estimates and
Census PEP estimates are not always consistent. In ACS estimation, smaller areas may be
combined and controlled as a single area for the purpose of population estimation. We
apply an additional factor of Census Bureau July-1 population divided by ACS 5-year
population for both the county and the city to resolve this discrepancy. This population
correction factor has the effect of controlling ACS estimates to the official Census
population estimates, as well as updating employment estimates to current vintage-year
Census population and allowing employment ratios to keep pace with population change.
Below is the formula for the ACS-share employment disaggregation ratio:

September 2015

LAUS Program Manual 10-2

Applying the ACS-share Employment Disaggregation
Procedure
Each year, data are updated to produce the ACS-share ratios as follows:

Developing City Employment/Population Index-Shares
ACS 5-year Estimates

Census July1 Est.

Employment
- Population
Correction

City

Employment

Population

Population

(I X III)/II

A
B
County Total

I
18,300
14,000
55,900

II
28,000
33,000
124,500

III
30,000
32,500
129,000

IV
19,607
13,788
57,920

ACS-share
ratio
IVCity 
IVCounty
V
0.338518
0.238048

Step 1. Data

from the most recent ACS 5-year data set on total employment and
population are entered in Columns I and II for all LAUS cities and city parts in the county.
Step 2. The

most recent July-1 population estimates are entered in Column III.

Step 3. For

each city a population correction factor is applied to its 5-year ACS
employment estimate. Column IV equals Column I times Column III divided by Column II.
Step 4. The

ACS-share employment disaggregation ratio is calculated for each city by
dividing the population-corrected employment level in the city by the population-corrected
employment level for the county.

September 2015

LAUS Program Manual 10-3

Claims-Based Unemployment Disaggregation
Research has shown that a strict claimant allocation method is not appropriate for
disaggregating total unemployment because local data, such as Unemployment Insurance
(UI) benefits claims, are not available for labor force entrant unemployed. Unlike most
job losers, the labor force entrant unemployed are usually not eligible for UI benefits;
they lack sufficient recent earnings due to employment. For this reason, claims data by
city or city part of residence are used to distribute only the experienced unemployed
component, i.e., those with recent job attachment.
Census population data are used in disaggregating unemployed new entrants and
reentrants. Population aged 16 to 19 data are used in the disaggregation of new entrants;
population data for those aged 20 and over are used in re-entrant disaggregation. Note
that these age groups are the same as those used to calculate the youth population ratio
for estimating county new entrant and reentrant unemployment using the Handbook
procedure.
Age-group population estimates for counties from the post-censal demographic series are
used for new entrant and reentrant unemployment disaggregation. Adjustments for outof-scope population are made using institutionalized group quarters and military group
quarters counts from the 2010 Census.
All new entrant and reentrant disaggregation ratios (R02 and R03, respectively) for cities
and city parts are calculated using the July-1 post-censal population estimates. The
calculations use population estimates from the latest post-censal vintage in conjunction
with counts by age group from the 2010 Census.

Required Claims Data for Claims-Based Unemployment
Disaggregation
The residency requirement for claims data is the coding and tabulating of claimants by
county of residence. The geographic distribution by residence of claimants filing
continued claims under State UI and UCFE certifying to unemployment in the week
including the 12th of the month is used to disaggregate the county estimate of
experienced unemployed to the city and city part level. Claimants with any earnings due
to employment in the week including the 12th should be excluded from counts used in
disaggregation. Though used for Handbook estimation, Railroad Retirement Board
(RRB) and commuter claims should be excluded from the claims counts used in
disaggregation.

September 2015

LAUS Program Manual 10-4

Claims-Based Unemployment Disaggregation Procedure and
Sequence
The procedure and sequence for claims-based unemployment disaggregation is presented
below, along with an example. The example assumes the following Handbook data:
• unemployment, excluding entrants (line 11)=
5,000
• reentrant unemployment (line 13)=
1,600
• new entrant unemployment (line 15)=
400
• total unemployment (line 16)=
7,000
• total county claimants without earnings =
4,500
• independent estimate of county unemployment(LAUS) = 12,000
City Claimants and Allocation Ratios
City

Claimants

>20 yrs.

16-19 yrs.

1
2

1,500
1,250

.25
.3

.2
.35

New entrant and reentrant disaggregation ratios represent 2010 Census ratios of city to
county age-group population counts adjusted based on population estimates from the
latest post-censal vintage.
Step 1. For a county, determine the percent of Handbook unemployment that is accounted
for by the experienced unemployed, those jobless with recent job attachment, i.e.,
unemployment excluding entrants divided by total unemployment.

Example: 5,000 ÷ 7,000 = 0.71
If any approved atypical adjustment was made to the UI data so that a claims count was
removed from the Handbook claims line leading up to Unemployment Excluding
Entrants, but is added to Total Unemployment, then that figure should be added to
Unemployment Excluding Entrants for purposes of arriving at the experienced
unemployed proportion.
Step 2. Determine

the proportion of county Handbook unemployment represented by
reentrants unemployment divided by total unemployment.

Example: 1,600 ÷ 7,000 = 0.23
Step 3. Determine

the proportion of county Handbook unemployment represented by new
entrants unemployment divided by total unemployment.

Example: 400 ÷ 7,000 = 0.06
Note: The proportions obtained in steps 1, 2, and 3 should sum to one (100%).
Example: 0.71 + 0.23 + 0.06 = 1

September 2015

LAUS Program Manual 10-5

Step 4. Apply

each of the proportions in steps 1, 2, and 3 to the independent county
estimate of total unemployed after additivity and adjustment to statewide controls. This
results in a disaggregation of total county unemployment into three parts:

A. experienced unemployed
B. reentrant unemployed
C. new entrant unemployed.
Example:

A = 0.7112,000 = 8,571.
B = 0.23 ,000 = 2,743
C = 0.06 ,000 = 686

Step 5. Allocate

the county estimate of experienced unemployed (estimate A in Step 4) to
all cities based on the percent distribution of place-of-residence claims data.

City

City
Claims

1

1,500

2

1,250



County
Claims

 4,500
 4,500

County
Exp
Unemp

City
Ratio

=
=

x
x

0.33
0.28

8,571
8,571

City
Exp
Unemp

=
=

2,857
2,381

Step 6. Allocate the county estimate of reentrant employment (estimate B in Step 4) to all
counties based on the percent distribution of the county’s population 20 years of age and
older from the 2010 census adjusted based on population estimates from the latest postcensal vintage.

City

County
Reentrants

1

2,743

2

2,743

20+ Pop
Ratios

x
x

City
Reentrants

=
=

25%
30%

686
823

Step 7. Allocate the county estimate of new entrant unemployment (estimate C in Step 4)
to all counties based on the percent distribution of the county’s population 16-19 years
old from the 2010 Census adjusted based on population estimates from the latest postcensal vintage.

September 2015

City

COUNTY
New
Entrants

1

686

2

686

16-19 Pop
Ratios

x
x

20%
35%

City
New Entrants

=
=

137
240

LAUS Program Manual 10-6

Step 8. Derive

the total unemployment estimate for each city by summing the city
estimates derived in Steps 5, 6, and 7.

September 2015

City

Step 5

1

2,857

2

2,381

Step 6

+
+

686
823

Step 7

+
+

137
240

unemployment

=
=

3,680
3,444

LAUS Program Manual 10-7

11Annual Processing
Introduction
n the current LAUS methodology, Handbook-based and model-based labor force
estimates are revised annually to take advantage of the latest available information.
This process is known as Annual Processing. State model performance is formally
reviewed by state, regional, and national office staff, and adjustments are made to
model specifications when necessary. Then new CPS population controls, revised
Handbook components, and revised state-supplied data are incorporated into the state and
substate estimates. In summary, annual processing consists of model evaluation and
performance review, incorporation of CPS population controls, collection and
incorporation of revised input data, re-estimation of state and substate estimates, and
benchmarking or additivity. The sections which follow discuss these processes in detail.

I

Annual Model Review
A benefit of using a model-based estimation framework is the ability to adapt a state's
model to the changing nature of the state economy and data. The variables in a model are
based on the inter-relationships in the state’s economy, including seasonal patterns and
long-term trends, and the individual nature of the data sources available. The variable
coefficients of the signal-plus-noise models allow the models to adjust gradually to
structural changes in the economy and to discount unusual changes of input data, such as
those resulting from CPS sampling variability. However, for some events, such as severe
weather or spurious movement in the CPS, it is important to be able to review a model’s
performance and take direct corrective action. In some cases, intervention variables are
added to the model to restore model performance; in other cases, model specifications are
revised. (See Chapter 6 for a detailed discussion of intervention variables and model
specifications.)
The LAUS model evaluation and performance review is conducted in the fall of each
year. First, a technical memorandum is issued which requests state staff to review their
model performance and provide comments and evaluations to the national office. The
memorandum usually includes a list of suggested topics on model behavior for the states

March 2016

LAUS Manual 11-0

to consider while reviewing their model performance. In addition, states are asked to
provide information about their economy which might help to explain model behavior.
At the same time, the Statistical Methods Staff (SMS) review the model performance
using statistical tests for diagnostic evaluation and outlier detection. This review focuses
on statistical measures of model performance and is in addition to the battery of statistical
tests which are run on the models each month. The tests help to determine whether any
changes to current model specifications or outlier interventions are necessary. SMS
shares the results of this research and their proposed actions with states, either via
technical memorandum or in presentations during the annual state/regional meetings.
LAUS national office staff monitor model performance as part of their monthly duties.
In addition, they summarize state issues and concerns regarding model performance and
present resolutions and answers during the annual state meetings. Questions and issues
raised by the states are also responded to in a formal LAUS technical memorandum,
generally issued just after the conclusion of annual processing activities.

Population Controls
At the beginning of each year, new CPS population controls are introduced for use in
division, state, and substate model estimation. These controls reflect both new data for
the most recent year and revisions to data for earlier years. The new and revised controls
are developed by the Census Bureau and delivered to BLS in late January.
Resident population at all levels of geography is estimated by updating a base population
from the latest decennial census via estimates of the components of population change,
consisting of births, deaths, and migration. The CPS universe, defined as the civilian
noninstitutional population, is then estimated by subtraction of the resident military and
institutional populations, primarily nursing homes, prisons and jails, mental hospitals, and
juvenile facilities. Below the national level, this procedure is supplemented by
estimations of domestic migration and by direct estimates of group quarters populations.
Where appropriate, estimates of student and military populations are also used.
Population controlling occurs when the sample-based monthly CPS labor force estimates
are adjusted so that they are consistent with these independently derived population
estimates. Adjusting (controlling) the CPS sample-based labor force estimates to be
consistent with independently derived population estimates reduces the variability of the
CPS estimates, thus improving their quality.
There are several ways CPS population controls affect LAUS estimates. For modelbased estimates, the monthly impact is via the CPS inputs to the model estimates. For
current estimation, monthly CPS population controls are incorporated into the CPS
estimates through the second-stage ratio adjustment step of the CPS estimation process.
(See Chapter 2 for a description of the second-stage estimation process.) For substate
estimates, the monthly impact of CPS population controls is less direct, and occurs due to
the additivity adjustment of substate estimates to their respective statewide or balance-ofstate totals.
Annual CPS population controls are incorporated into CPS estimates through the revision
of monthly CPS labor force estimates at the end of the year. During annual processing,

March 2016

LAUS Manual 11-1

the revised CPS data are incorporated into the LAUS estimates when the models are reestimated.

How Population Estimates are Calculated

Current estimates of the national population by age, sex, race, and Hispanic origin are
derived each year by taking the decennial base population and adjusting it forward
through time using components of population change. Census produces estimates using a
component cohort method, derived from the following demographic balancing equation.

That is, the population at any given time point is the population at the last decennial
Census, plus all U.S. births, less U.S. resident deaths, and adjusted for migration. While
the equation is the same for all areas and demographic groups, differences in data
availability dictate that no single methodology is capable of producing estimates for all
areas and characteristic groupings.
Geographic and Demographic Consistency: Census creates population estimates via a
‘top-down’ approach. First, national estimates of monthly population are created by age,
sex, race, and Hispanic origin, as estimates of the components of change are more
reliable at the national level. Next, annual population estimates by demographic groups
are created for counties. Annual state estimates are created by aggregating the county
values. The national monthly population time series are then used to create monthly state
and county estimates by age, race, sex and Hispanic origin in such a way that ensures
geographic and demographic consistency in all areas.
County total population counts are created using a simple control, so that the sum of
county total populations equals the national total. Demographic characteristics estimates
are controlled via a two-way raking system. This ensures that the aggregation of
subnational demographic population estimates sum to the national demographic
population estimates (e.g., the sum of all county Hispanic population estimates must
equal the national estimate). Simultaneously, the sum of different demographic
population group estimates is also controlled to the total population for the state or county
(for example, the sum of male and female population in a state must equal its total
population).
Enumerated Base Population: The base for all postcensal population estimates begins
with the most recent decennial census enumerated counts. These counts are modified in
three ways to produce the April 1, 2010 population base.
1. The Count Question Resolution (CQR) program allows legal entities to challenge
their jurisdiction’s decennial census value. If this challenge is successful, it is
incorporated into the base population.
2. Changes to legal boundaries reported by January 1 of the vintage year are used to
create a population base and a full time series consistent with those new legal
boundaries. This generally does not impact national or state estimates, but can
impact counties, cities or towns.
March 2016

LAUS Manual 11-2

3. Race categories are modified to be consistent with the race categories from available
data input sources. The 2010 Census allowed for responses that included one or
more race groups, as defined by the Office or Management and Budget (OMB). It
also allowed for responses of “Some other race”. If “Some other race” was chosen in
combination with other races, the “Some other race” component was simply
removed. If it was chosen alone, then the response was allocated to one of the other
race categories using other household information if available. If not, a hot-decking
procedure is used instead.
Vital Statistics (Births and Deaths): Vital statistic (birth and death) data come from two
sources: The National Center for Health Statistics (NCHS) and the Federal-State
Cooperative for Population Estimates (FSCPE). NCHS data are derived from birth and
death certificates across the United States. The FSCPE contributes data on the
geographic distribution of vital events within the state. Census also computes short-term
projections of vital events to account for the lag in availability of vital statistics.
Adjustments to birth data are made to account for discrepancies between state race data
categories and those of OMB, birth certificates that include race or Hispanic origin of
parents but not of the child, and inconsistencies between imputed race distributions and
those reported in the Census enumerated counts.
Modifications are also made to deaths data. Many states still use the 1977 OMB race
categories. These are converted to the 1997 categories. In addition, Census accounts for
the less reliable age-of-death reporting for person 70 years of age or older by
redistributing all deaths occurring to the aggregate population “70 year and older” by sex,
race, and Hispanic origin to a single year of life (70 to 99) or to 100+ years using lifetable-based death rates.
For subnational estimates, direct data are used where available. Where data are not yet
available, short-term projections are created by calculating county-level age-specific
fertility and mortality rates. These rates are applied to the previous vintages population
projections and reconciled with FSCPE data on the geographic distribution of total
county vital events. These data are then controlled using the method described in a
previous section, “Geographic and Demographic Consistency.”
Net Migration (International and Domestic): Migration is the third major component of
the balancing equation. For national population estimates, domestic migration (between
county and/or state) nets to zero, so only international migration is used. For subnational
estimates, both international and domestic migration are important components of
change.
International migration, in its simplest form, is any change of residence across the
borders of the United States. Census divides international immigration into (1)
immigration of the foreign-born; (2) emigration of the foreign-born; (3) net migration
between the United States and Puerto Rico; (4) net migration of natives to and from the
United States; and (5) net movement of the Armed Forces population to and from the
United States.
1. Immigration of the foreign-born is estimated separately for Mexico and “All other
countries” using American Community Survey (ACS) data on residence one year ago
March 2016

LAUS Manual 11-3

(ROYA). Adjustments are made to account for children less than one year of age (of
whom the ROYA question is not asked). For county estimates, ACS Year of Entry
(YOE) data are used to distribute the national total foreign-born immigration data by
geographic and demographic detail. Responses with a YOE within the last five years
are used as a proxy to distribute the national data to state and county characteristic
population estimates for the foreign-born.
2. Emigration of the foreign-born is calculated by using ACS data on nine potential
emigration groups. For each group, the associated group’s population from one 5year ACS estimate is aged forward using NCHS life tables. This creates an expected
population at a later points in the same 5-year file. Six residuals are created by
subtracting the estimated population from the actual population (three 2-year
residuals, two 3-year residuals, and one 4-year residual). These are averaged together
to created estimated emigration rates with reduced variability. These rates are
applied to 1-year ACS data to obtain annual estimates of foreign-born emigration.
3. Migration between the United States and Puerto Rico is estimated using data from
both the ACS and the Puerto Rico Community Survey (PRCS). Immigrants are
derived from persons in the ACS reporting residence in Puerto Rico one year prior.
Emigrants are persons in the PRCS reporting residence in the United States one year
prior.
4. Migration of native-born migrants is created using census and population register
data from over 80 different countries. Estimates of U.S. natives residing in each
country are compared between two consecutive years. The difference between the
two year’s data is used to develop an average annual estimate of net native-born
migration.
5. International movement of Armed Forces population is estimated from data collected
by the Defense Manpower Data Center (DMDC). DMDC provides monthly
tabulations of military personnel station or deployed outside of the United States, by
demographic group and service branch. Changes in overseas military population,
excluding deaths, is assumed to be movement of personnel into and out of the United
States. County data from the DMDC is used to estimate net international movement
at the county-level, with data from the most recent ACS five-year file used to
improve the geographic distribution around some domestic military installations.
Domestic migration has no impact on national estimates, but is an important
consideration in subnational population estimates. County-to-county net domestic
migration (NDM) is based on data from three sources: Internal Revenue Service (IRS)
tax exemptions, change in Medicare enrollment, and changes in the group quarters
population.
NDM data are produced for three age groups: under 18, 18 to 64, and 65 plus. Address
data from IRS tax returns for individual filers are compared to produce geographic data
by age categories for the under 18 and 18-to-64 age groups. Not all persons are tax filers,
so the level data cannot be used. Rather, these data are used to create NDM ratios for
each age breakout.
IRS data coverage is lower for persons age 65 and over, so changes in Medicare
enrollment from the Centers for Medicare and Medicaid (CMS) are used to calculate
March 2016

LAUS Manual 11-4

migration ratios for this age group. Again, levels cannot be used, as not all persons are
eligible and/or enrolled. Rather, an NDM rate is calculated by taking change in Medicare
enrollment (less persons turning 65 (newly eligible), deaths, and net international
migration) and dividing it by total Medicare enrollment at the start of the period.
For each age group, the calculated NDM rate is applied to household population
estimates for the age group and area. This creates estimates of total county net domestic
migration, as well as by demographic group. The county data are then controlled so that
total domestic migration at the national level is zero.
Data for state and county estimates by demographic characteristics come primarily from a
combination of IRS tax exemptions, the Social Security Numeric Identification File
(NUMIDENT), and the Person Demographic Characteristics File (PDCF). The
NUMIDENT provides information on age and sex. The PDCF is derived from previous
decennial censuses and a variety of administrative records sources, and it is used to
estimate the race and Hispanic origin of the exemptions. This gives estimates of inmovers and out-movers by characteristic.
To account for under-coverage in IRS data, ratios are computed for outgoing migration
for each demographic grouping. Total out-movers are created by applying this ratio to
the population estimates for each demographic group within a county. The sum of outmovers is then also distributed as county in-movers using the incoming county migration
data. Because in-movers are not calculated directly from the population estimates, but
rather by sharing out out-movers, total national domestic migration by demographic
grouping always nets to zero; that is, each and every in-mover is also an out-mover
somewhere else, and vice versa.
Civilian Non-Institutional Population: The universe for BLS household estimates in the
Civilian Non-Institutional Population, ages 16 and older. To create population estimates
to serve as controls for surveys such as the CPS, the total population estimates are
adjusted to remove armed forces personnel, the institutionalized population (from
estimates of non-military Group Quarters population), and persons under the age of 16.
For more information on the development of national and state population estimates, see
the US Census Bureau documentation at
http://www.census.gov/popest/methodology/2015-natstcopr-meth.pdf.

March 2016

LAUS Manual 11-5

State Annual Population Controls

Each January, the Census Bureau provides BLS with revised population estimates for
each state. There are three types of revisions that may be incorporated into the revised
population estimates:
1. Base Population Updates: The most recent decennial base population may be
updated by incorporating new Count Question Resolution (CQR) changes, as well as
legal boundary changes and other geographic updates.
2. Changes to input data and methods of estimating components of change: The data
used to estimate the individual components of population change are updated as more
recent and/or complete data become available. In addition, Census may revise the
actual methods of estimating the components of population change and/or the stock
estimates for special populations. Generally, revisions to the inputs tend to affect
estimates toward the end of the time series. Revisions to methodology often affect
the series cumulatively from the data of the last Census and forward.
3. Changes to the method of estimating population: In addition to updating the method
of estimating the components of population change, the Census Bureau may revise
the method of estimating population using the components of change. This type of
revision also affects the series from the census date forward.
Once BLS receives the population controls from the Census Bureau, they are used to
adjust the state's CPS labor force data. This is done by ratio adjustment: each month’s
CPS employment and unemployment estimates is multiplied by the ratio of that month’s
revised population value divided by the original population value.
Under normal circumstances, the CPS unemployment rate is not affected by population
re-controlling. As both the unemployment estimate (the numerator of the rate
calculation) and the labor force estimate (the denominator of the rate calculation) have
been adjusted by the same proportion, the ratio of the two is unchanged.

Annual Population Revisions Affecting Substate Area Estimates
Substate area estimates benefit from the monthly and annual CPS population revisions
through the additivity process which assures consistency with the state totals. Substate
estimates utilize county and place total population data in the disaggregation of labor
market area estimates into smaller geographic entities. Total population estimates for
counties, incorporated places, and minor civil divisions are revised by the Census Bureau
each year.

March 2016

LAUS Manual 11-6

Annual Re-Estimation
Each year, states are provided the schedule for annual processing activities in a technical
memorandum. States are instructed to replace the model input data with revised Current
Employment Statistics (CES) employment, striker, unemployment insurance (UI)
claimant, and Unemployment Compensation for Federal Employees (UCFE) claimant
data for every period for which they have revisions or corrections. The CES nonfarm
wage and salary estimates should reflect the most recent Quarterly Census of
Employment and Wage (QCEW) benchmark and include any changes beyond the regular
two years of CES benchmarking. Claims counts should be updated wherever possible.
Model-based annual processing is done in multiple stages. Division-model outliers are
simply the sum of state outliers, so all state outliers must be designated prior to any
annual processing re-estimation. Outliers are rarely impacted by input revisions (either to
the CPS via population controls or to the covariate CES or continued claims series).
Therefore, calculation of outlier effects is done prior to the full run of annual processing.
SMS re-specifies models and outliers in each series immediately following the
completion of statewide estimation for December of the previous production year. Their
work is based on their own review and that of each state partner (see earlier section
“Annual Model Review”).
Once models are re-specified, full annual processing operations may commence.
Updated population controls from the Census Bureau are verified and applied to the CPS
employment and unemployment series. These new estimates are used to update the
inputs to variance and error calculations. LAUS division models are then re-estimated.
During this time period, the STARS web interface is placed into Annual Processing
mode. Access to monthly production tools, including the data extract utility, are
unavailable. States use STARS to provide revised input data to the program office. State
groupings for model-based annual processing are on a Census division basis because realtime benchmarking is performed at the division level. Once all states in a division have
entered their input data and the national office has validated them, the historical series for
all states in the division are re-estimated and benchmarked. Typically, five years of
model-based estimates are revised during annual processing.
Observations in time-series models borrow strength from other time periods. In reestimation, the entire time series is used to re-estimate every observation. The estimation
process is run forward from the beginning of the time series, run backward so that earlier
observations benefit from later data, and then run forward again to the end of the year.
This is possible because LAUS models use a Forward Filter to modify each of the
model’s coefficients with the addition of each monthly observation.
The Forward Filter acts as weighting mechanism, which allocates how much a model’s
coefficients will change (and thus the estimates) with each new period’s data. Because
the Forward Filter works by evaluating each successive observation, one after another,
the models can produce estimates both forward and backward through time. During the
initial forward pass, each successive estimate incorporates all of the information from the
earlier months in the time series. At the end of the time series, the estimation process is
performed backward through time, so that each past month’s estimate can benefit from
the more recent data. Finally, the process is performed moving forward through time

March 2016

LAUS Manual 11-7

again, so that information from the first two passes can be incorporated into the entire
time series.
As with monthly estimation, annually processed estimates are subject to real-time
benchmarking. An estimation criterion is added to the estimating procedure, so that the
sum of the not seasonally adjusted (NSA) estimates are forced to sum to those of a higher
level of geography. Divisions are benchmarked to the national CPS total (including
population re-controls, which means this does not exactly match national estimates as
published by the BLS CPS program). States are controlled to their respective Census
division benchmarked estimates, again via estimation criterion to allocate the benchmark
discrepancy. Modeled areas and balances-of-state are controlled to their respective states
via pro-rata ratio adjustment.
The ratio for benchmarking NSA estimates is applied to the seasonally-adjusted (SA)
series. These benchmarked SA series are smoothed using the historical, two-sided Trend
Cycle Cascade Filter (TCCF) smoother. This filter is symmetric in the middle of the time
series, but grows increasingly asymmetric at the ends, as there are insufficient
observations to maintain full symmetry. December of the most recent production year
has no subsequent observations to incorporate into the TCCF, so its filter is identical to
the current production filter. This method eliminates methodological discontinuities
between December and January estimates.
Once the annually processed data have been verified and the program office is reasonably
confident that no further input revisions are forthcoming, annual processing mode is
turned off for the completed division(s). At this point, completed states are notified that
annual processing has finished and STARS is now fully available. State users may then
use the extract utility to extract the revised LAUS estimates, as well as run normal
monthly processing activities for January of the new production year. These include
viewing national office simulations, running state simulations, and transmitting January
inputs for production. If any questions arise about the annually processed estimates, they
should be directed to the appropriate BLS regional office.

March 2016

LAUS Manual 11-8

12 STARS Output Tables
Introduction

A

ll states use the monthly web based system called STARS (State Time
Series Analysis and Review System) each month to produce their
statewide labor force estimates, and to transmit these to the BLS national office.
In producing estimates, States provide input data such as CES employment,
strikers, and unemployment insurance claims. These data are used not only to
produce the estimates, but are also stored in a national office database and are
available through STARS, along with data from the CPS, for use in various
analytical studies and model interpretation. Each time STARS is run, it provides
both BLS and State analysts with output containing a series of tables and graphs
with information for examining employment trends, preparing releases, or
understanding the nature of a month-to-month change in model estimates.
The primary functions of STARS are to:
• integrate State/BLS data entry.
• calculate monthly state level estimates for the current month and revised
estimates for the previous month for labor force, employment, unemployment,
and the unemployment rate.
• provide error measures, analytical charts and tables.
CPS data are loaded into STARS at the national office as soon as the monthly
national press release is issued. State analysts can enter their inputs when they
become available, review their listings, and once they are verified, final LAUS
estimates are produced by the LAUS national office through the STARS system.
States also have the option to run estimates with preliminary numbers before the
actual data are available without transmitting the estimates to BLS. This option
available using the State Simulation module under the Monthly Processing menu.
A STARS User’s Guide is available to assist users in creating and updating their
monthly model-based estimates using the web-based STARS interface. The

LAUS Program Manual 12-1

user’s guide is designed to introduce new users to the STARS interface and
provide them with the basic skills required to operating STARS.
The latest version of the STARS User’s Guide is available at the STARS website
under the Help link on the login screen menu. Users are not required to log into
STARS to access it. The user’s guide can be viewed online, printed, or
downloaded by individual chapter or in its entirety in PDF format. The guide can
also be referenced while a user is logged into the system.

STARS Review Estimates
The Review Estimates link in the STARS Monthly Processing Menu
enables you to review current, finalized estimation output tables, as well
as archived historical estimation output tables for the selected State, substate areas (i.e., metropolitan areas, balance of State) and Census
Division. Historical monthly estimation output is available from January
1978 through the month for which the most recent estimates have been
processed. You can view the output tables online and print the tables on
request. (See Chapter 3, pages 23-29 of the STARS User’s Guide.)
The estimation output tables for the specified month and year will be
displayed. Output tables are available at the State level, the area model level and
the Census Division level.
Below is the header page that precedes the estimation output tables. It provides
basic information about the estimation run. Starting at the top, it displays the
reference month/year and the State. Then it lists the date and time for when the
estimation was run. Next it provides a quick look at the inputs that were entered
for the current and previous months.

LAUS Program Manual 12-2

State

Following the header page are 31 tables and 8 sets of charts. Explanations and
examples of the tables and figures are provided on the following pages.

LAUS Program Manual 12-3

STARS Table 1: Year-to-Date Model Estimates
Table 1 provides all the year-to-date LAUS model estimates for labor force,
employment, unemployment, and the unemployment rate. A quick comparison
of current estimates to earlier estimates within the same calendar year can be
made. Tables 1a and 1b indicate when the monthly changes are significant at the
5 percent level (**) and the 10 percent level (*). The error range for the
unemployment rate at the 90 percent confidence interval is also displayed.
Tables 1c and 1d show the over-the-month changes in the smoothed seasonally
adjusted and not seasonally adjusted series. Developing trends and month-tomonth changes can be observed. A comparison of the seasonally adjusted and not
seasonally adjusted series can be made.

** Significant change at 5% level
* Significant change at 10% level
+ 90% Confidence Interval

LAUS Program Manual 12-4

** Significant change at 5% level
* Significant change at 10% level
+ 90% Confidence Interval

LAUS Program Manual 12-5

STARS Table 2: Standard Errors for Year-to-Date Model
Estimates
This table shows the monthly standard errors for each of the labor force
components displayed in Table 1. The standard error refers to the variability of
an estimate and is used in the construction of confidence intervals.

LAUS Program Manual 12-6

STARS Table 3: Over-the-Year Changes
This table shows over-the-year changes and standard errors for the each of the
basic types of labor force estimates. The level of change and the percent change
are given in Tables 3a and 3b for the smoothed seasonally adjusted and not
seasonally adjusted series respectively. Tables 3c and 3d show the standard errors
for the over-the-year changes in the two series. The data for all years prior to the
current year reflect the annual updating. The over-the-year changes give an
indication of the state's labor force trends.

LAUS Program Manual 12-7

Tables 4-6: Components of Change
These tables show the components of the unemployment rate, unemployment and
employment of both the model and the CPS and how they changed over the
month.
Table 4 displays the components of change for the unemployment rate. Table 4a
contains the level, trend and seasonal change for the model. It also includes
components of change for the two inputs to the model. Table 4b contains the
changes for the CPS, the signal and the noise. The same items are shown for the
unemployment level in Table 5 and the employment level in Table 6.
Since each model estimate is the sum of its variable components, the analyst can
see the influence that each variable has on the total estimate by examining the
components of change. It may be that one input variable is the primary influence
in the current over-the-month change in the estimate.
The analyst can determine if the variable components are behaving in their
"normal" way for this time of year by comparison with last year’s behavior.

** Significant change at 5% level
* Significant change at 10% level

LAUS Program Manual 12-8

** Significant change at 5% level
* Significant change at 10% level

** Significant change at 5% level
* Significant change at 10% level

LAUS Program Manual 12-9

Tables 7-8: Standard Errors
Table 7 shows the standard error for the model components of change for the not
seasonally adjusted, seasonally adjusted, and smoothed seasonally adjusted
series.

Table 8a lists the monthly CPS estimates and table 8b shows the standard
error for the CPS levels and changes.

LAUS Program Manual 12-10

Tables 9-10: State Data, Trend and Seasonal Factors
Table 9 displays the seasonal factors that are applied to the unemployment
rate, the unemployment level, the employment level and the employmentpopulation ratio to create seasonally adjusted estimates.

In table 10a the levels, trends and seasonal factors of the UI claims and
CES inputs are exhibited.

In table 10b the regression coefficients for the UI claims and CES inputs
are shown for both benchmark and not benchmarked estimates.

LAUS Program Manual 12-11

Table 11: Diagnostics, Prediction Error and State Inputs
This table is useful for checking new data for unusual values. Before
receiving new data into the model, values are predicted from the
accumulated historical experience of the State inputs and CPS data.
Observations that are far from the predicted value can identify outliers,
which are unusually large changes; inliers, which are unusually small
changes; or incorrect data due to mistakes.

LAUS Program Manual 12-12

Table 12: Benchmark Adjustment Factors
State levels are benchmarked to the Division totals. Table 12 displays the
State model adjustment for signal and trend applied for unemployment and
employment.

LAUS Program Manual 12-13

Tables A-1, A-2, and A-3: Gen4 Monthly Analysis of Change
Tables
These tables provide information on the monthly change of the official
estimates. Table A-1 shows the unemployment rate series, A-2 shows the
unemployment series, and A-3 shows the employment series. The effect of
real-time benchmarking is shown for both the not seasonally adjusted and
seasonally adjusted estimates. The directional persistence of the model
from smoothing is also displayed here.

LAUS Program Manual 12-14

Figures 1-8
Also included in the STARS output are figures that visually display the
input data and seasonal factors.
Figure 1 Unemployment Rate
Figure 1a charts the seasonally adjusted LAUS unemployment rate and the
claims rates. Figure 1b displays the unadjusted LAUS and CPS
unemployment rates with the claims rates.
Figure 2 Unemployment
Figure 2a charts the seasonally adjusted LAUS unemployment level and
the claims level. Figure 2b displays the unadjusted LAUS and CPS
unemployment levels and the claims level.
Figure 3 Employment
Figure 3a charts the seasonally adjusted LAUS employment and the CES
employment level. Figure 3b displays the unadjusted LAUS and CPS
employment and the CES employment level.
Figures 4-7 Unemployment Seasonal Factors
Seasonal factors indicate the expected seasonal variation in the series.
Often differences in the seasonal patterns help to explain the difference
between the CPS and State inputs series in their direction and magnitude
of change.
Figure 4a shows the seasonal factors and the seasonal means for the not
benchmarked LAUS unemployment series. Figure 5a shows the seasonal
factors and the seasonal means for the benchmarked LAUS unemployment
series. Figures 4b and 5b show the seasonal factors and the seasonal
means for the claims series.
Figure 6a shows the seasonal factors and the seasonal means for the not
benchmarked LAUS employment series. Figure 7a shows the seasonal
factors and the seasonal means for the benchmarked LAUS employment
series. Figures 6b and 7b show the seasonal factors and the seasonal
means for the CES employment series.
Figure 8 CPS Population
Figure 8 charts the CPS population estimate for the State.

LAUS Program Manual 12-15

Figure 1: Unemployment Rate
State

State

LAUS Program Manual 12-16

Figure 2: Unemployment Level
State

State

LAUS Program Manual 12-17

Figure 3: Employment Level
State

State

LAUS Program Manual 12-18

Figure 4-5: Unemployment Seasonal Factors
State

State

State

State

LAUS Program Manual 12-19

Figure 6-7: Employment Seasonal Factors
State

State

State

State

LAUS Program Manual 12-20

Figure 8: CPS Population Estimate

LAUS Program Manual 12-21

LAUS
Glossary

Glossary
Additional Benefits: (AB) See State Additional Benefits.

Additional Claim: A notice of new unemployment filed at the beginning of a second or subsequent series
of claims within a benefit year or within a period of eligibility when there has been intervening
employment. This is one of three types of initial claims.

Additivity Adjustment: The procedure which forces the exhaustive Handbook estimates to equal the State
estimate is known as additivity adjustment. The process is usually linear unless an atypical procedure is in
effect. The linear additivity adjustment is accomplished through the Handbook share procedure of linking
counties and MCDs to the CPS-based State estimate.

Agent State: The State in which a claimant files an interstate claim for compensation against another
(liable) State where wages were earned is the agent State. Usually, this is the claimant’s State of
residence.

All Other Nonagricultural Employment: This includes self-employed, unpaid family workers, and
domestics in private households.

American Community Survey (ACS): A household survey developed by the Census Bureau to replace the
long form of the decennial census program. The ACS is a large demographic survey collected throughout
the year using mailed questionnaires, internet questionnaires, telephone interviews, and visits from
Census Bureau field representatives to about 3.5 million household addresses annually.

Annual Processing (AP): A series of activities conducted annually which result in benchmarked State and
substate estimates. These activities include the State submission of revisions to model inputs, revisions to
substate inputs, incorporation of revised population controls, model re-estimation and smoothing,
benchmarking, and seasonally adjusting the revised series.

Areas of Substantial Unemployment (ASU): This is defined under the Workforce Investment Act as an
area of at least 10,000 population with an average of 6.5 percent unemployment or higher in the previous
12 months. It is used for determining eligibility for employment and training programs.

Auto Regressive Integrated Moving Average (ARIMA): A statistical approach designed to make forecasts
of a time series based on only its past values. Part of the non-model-based X-11, including X-11 ARIMA
and X-12 ARIMA, which has been the standard BLS approach to seasonal adjustment since the 1970s.

Autocorrelation: Identifies whether the error terms in a regression equation are not independent over time.
If this is not accounted for in the equation for the regression line, poor coefficients and predicted values
may result. All State models have coefficients adjusted to reflect autocorrelation.

Autocorrelation Coefficient or  (rho): A mathematically determined value that measures the relationship
or correlation between successive error terms of the same series. A value of "0" means that there is no
correlation and a value of "1" indicates total positive autocorrelation.

Base Period (Base Year): A base period is a specified period of twelve consecutive months (or in some
States, 52 weeks preceding the beginning of a benefit year) during which an individual must have the
required employment and/or wages in order to establish entitlement to compensation or allowances under
an applicable program.

Benchmark: This is a point of reference (either an estimate or a count) from which measurement can be
made or upon which adjustments are based.

Benefit Year: A period, generally a 52-week period, during which individual claimants may receive their
maximum potential benefit UI amount.

Bias: The difference between the expected value of the estimate from a probability sample and the true
value of the population parameter.

The Bureau of Labor Statistics (BLS): Is the principal fact-finding agency for the Federal Government in
the broad field of labor economics and statistics. The BLS is an independent national statistical agency
that collects, processes, analyzes, and disseminates essential statistical data to the American public, the
U.S. Congress, other Federal agencies, State and local governments, business, and labor. The BLS also
serves as a statistical resource to the Department of Labor. BLS data must satisfy a number of criteria,
including relevance to current social and economic issues, timeliness in reflecting today’s rapidly
changing economic conditions, accuracy and consistently high statistical quality, and impartiality in both
subject matter and presentation.

Bureau of the Census: A bureau of the U.S. Department of Commerce that serves as leading source of
data about the nation’s people and economy. The primary mission of the Bureau is to conduct censuses of
population and housing every 10 years and of agriculture, business, governments, manufacturers, mineral

industries, and transportation at five year intervals. It also conducts the America Community Survey
(ACS) as well as the Current Population Survey (CPS) for the BLS.

Census: A count or enumeration (as opposed to a sample or an estimate) of a specified population or some
other characteristics in a given area (housing, industry, etc.)

Census Share: A method formerly used to disaggregate LMA employment and unemployment estimates
to smaller areas by assigning the same proportion of the monthly, independent LMA estimate evidenced
in the most recent census.

Census Tracts: Small, relatively permanent statistical subdivisions of a county that provide comparable
population and housing census tabulations. Tracts are designed to be relatively similar in population
characteristics, economic status, and living conditions. The average tract has about 4,000 inhabitants.
Census tract boundaries are recommended by local census tract committees and approved by the Bureau
of the Census.

Certification (Certifying): The process and form by which a claimant states and attests to facts which will
determine eligibility for UI benefits for a given week. These include, for example, a search for work and
availability for work.

Civilian Noninstitutional Labor Force: The sum of all employed and unemployed persons excluding
persons under 16 years of age, inmates of institutions, and members of the Armed Forces.

Claim: A notice of unemployment filed by an individual to request a determination of unemployment
insurance eligibility and the amount of benefit entitlement, or to claim benefits or waiting-period credit.

Claimant: A person who files either an initial claim or a continued claim under (1) any State or Federal
unemployment compensation program or (2) any other program administered by the State agency.

Claims-Based Unemployment Disaggregation: A method for disaggregating LMA unemployment to
subareas by using (1) claims by county of residence to distribute Handbook experienced unemployment
and (2) CPS-based data to allocate Handbook new and reentrant unemployment. It is used in conjunction
with the population-based indexed share employment disaggregation.

Class of Worker: There are three classes of workers: (1) wage and salary workers who receive wages,
salary, commission, tips, or pay in kind from an employer; (2) self-employed persons who work for profit

or fees in their own business, profession, or trade, or on their own farms; and (3) unpaid family workers
who work without pay for 15 or more hours a week on a farm or in a business operated by a household
member to whom they are related by birth, marriage, or adoption.

Coefficients: The values of the intercept and slope in the formula for the regression line. Coefficients are
estimated by a mathematical formulation which calculates coefficients by minimizing the squares of the
differences between the actual values (Y) and the predicted values (Y'). They represent (mathematically)
the relationship of the independent variable to the dependent variable and how the changes in one variable
can be related to another. In the case by a of a variable coefficient model, the coefficients are allowed to
change over time to reflect changes that are occurring in the relationships of the dependent and the
independent variables.

Coefficient of Variation (CV): The measure of relative dispersion of data. The standard deviation divided
by the arithmetic mean times 100 yields the coefficient of variation.

Combined Statistical Area: A geographic entity consisting of two or more adjacent Core Based Statistical
Areas (CBSA) linked through commuting ties.

Commutation: Regular travel of a person from the place of residence to the job location or to the place of
filing for UI benefits.

Commuter Claimant: Under the Intrastate Benefit Payment plan, a worker who travels regularly across a
State line from home to work, and by mutual agreement between States, files in the State where the
individual last worked when employed, and is treated as a resident of that State.

Compositing: An estimating technique which combines information from different sources, taking into
account the relative accuracy of each source. In the LAUS regression models, the Kalman Filter technique
can be thought of as a type of compositing. It combines CPS and model estimates using their variances as
a measure of the accuracy of the data.

Continued Claim: A claim filed after the initial claim, by mail, internet, telephone, or in person, for
waiting-period credit or payment for a certified week of unemployment.

Core: A densely settled concentration of population, comprising either an urbanized area (of 50,000 or
more population) or an urban cluster (of 10,000 to 49,999 population) defined by the Census Bureau,
around which a Core Based Statistical Area is defined.

Core Based Statistical Area (CBSA): A statistical geographic entity consisting of the county or counties
associated with at least one core (urbanized area or urban cluster) of at least 10,000 population, plus
adjacent counties having a high degree of social and economic integration with the core as measured
through commuting ties with the counties containing the core. Metropolitan and Micropolitan Statistical
Areas are the two categories of Core Based Statistical Areas.

Correlation: A statistical term which indicates a structural, functional, qualitative correspondence between
comparable entities. Correlation can be either positive (simultaneous increase or decrease in both
variables) or in negative (increase in the value of one and decrease in the value of the other variable).

Correlation Coefficient: A mathematically determined value that measures the relationship or correlation
between two time series. As with the autocorrelation coefficient, a value of "0" indicates no correlation
and a value of "1" indicates a total positive correlation. A value of "-1" indicates total negative
correlation, or meaning that as one series increases, the other series decreases.

Covered Employment: Those jobs covered by the unemployment compensation programs are considered
covered employment. At this time, those not covered include some agricultural workers, employees of
religious and small nonprofit organizations, household workers, and self-employed workers.

Current Employment Statistics (CES): A BLS monthly survey of about 144,000 businesses and
government agencies, representing approximately 554,000 individual worksites that yields estimates of
nonagricultural wage and salary employment, hours, and earnings by industry. These statistics are
prepared monthly for the nation as a whole, and by cooperating State agencies for all 50 States, the
District of Columbia, Puerto Rico, the Virgin Islands, and about 400 metropolitan areas and divisions.

Current Population Survey (CPS): A monthly survey conducted by the Bureau of the Census of
approximately 60,000 assigned households of which 50,000 are eligible for interview. This survey of the
civilian noninstitutional population of the United States provides monthly statistics on employment,
unemployment, demographic characteristics, and related subjects which are analyzed by the Bureau of
Labor Statistics.

Denial of Benefits: An action imposed by a State agency after a nonmonetary determination or an appeals
decision which cancels, reduces, or postpones a claimant's benefit rights.

Dependent Variable: The variable for which estimates are desired, usually termed the "Y" variable. In the
LAUS models, the dependent variable used in constructing the model is the monthly CPS estimate.

Determination: An official decision by the State UI agency regarding the unemployment claim of a
person. (See monetary and nonmonetary determination.)

Directional Persistence (DP): A measurement of the strength of current movement in a smoothed
seasonally adjusted (SSA) estimate. Directional Persistence indicates how much the seasonally adjusted
(SA) estimate must change in the next period to reverse the direction of change in SSA estimate. If the
SSA estimate is declining then the change in SA estimate for the next month must be greater than the DP
value to increase the SSA. If SSA estimate is increasing then the change in SA estimate for the following
month must be less the DP value to decrease the SSA estimate.

Disaggregation: A method to divide a statistic into its component parts. For example, the LMA
unemployment is divided into each component county or city.

Disaster Unemployment Assistance (DUA): A program that provides unemployment assistance to
individuals whose unemployment is a direct result of a major disaster as declared by the President of the
United States.

Discouraged Workers: Individuals not in the labor force who want and are available for a job and who
have looked for work sometime in the past 12 months (or since the end of their last job if they held one
within the past 12 months), but are not currently looking, because they believe there are no jobs available
or there are none for which they would qualify.

Dynamic Residency Adjustment Ratios (DRR): A method to adjust CES employment data for resident
employment in an area that accounts for the relationship between employed residents and jobs in that area
and in other areas within commuting distance and job growth within these areas.

Earnings Disregarded: The amount prescribed by State unemployment compensation laws that a claimant
may earn without any reduction in weekly benefit amount for a week of total unemployment. This is also
referred to as the forgiveness level for earnings. The amounts vary by State.

Earnings Due to Employment: These are earnings, either from the regular employer or from odd jobs,
which a UI claimant may receive while certifying to a week of unemployment. The existence of these
earnings classified the claimant as employed, even when earnings are less than the State's forgiveness
level.

Employed: In the CPS, those individuals 16 years of age or older who worked at least one hour for pay or
profit or worked at least 15 unpaid hours in a family business during the reference week are considered
employed. Individuals are also counted as employed if they had a job but did not work because they were

ill, on vacation, in a labor dispute, prevented from working because of bad weather, or taking time off for
personal reasons.

Employment/Population Ratio: The proportion of the civilian noninstitutional population who are
classified as employed.

Employment and Training Administration (ETA): Agency under the Department of Labor that
administers federal government job training and worker dislocation programs, federal grants to states for
public employment service programs, and unemployment insurance benefits. These services are primarily
provided through state and local workforce development systems.

Enumeration Districts (EDs): Administrative units used in the Census. They contain, on the average,
about 750 people. The EDs provide a list of addresses for housing units which is used to help set up the
sample file for the CPS.

Error: See Standard Error.

Establishment: An economic unit which produces goods or services, is generally found at a single
physical location, and is primarily engaged in one type of economic activity.

Exhaustees: Individuals who have exhausted all of the unemployment insurance benefits to which they
are entitled within a benefit year and cannot establish a new benefit year.

Extended Benefits (EB): The supplemental program, established by Public Law 91-373 that pays
extended compensation during a period of specified high unemployment to individuals for weeks of
unemployment after they have exhausted regular compensation. The program is financed equally from
Federal and State funds and becomes operative at the State level. The State determines benefits and
certain restrictions.

Extended mass layoff event: A layoff defined by the filing of 50 or more initial claims for unemployment
insurance benefits from an employer during a 5-week period, with at least 50 workers separated for more
than 30 days. Such layoffs involve both persons subject to recall and those who are terminated.
Extrapolation: A method to project values of a variable in an unobserved interval from values within an
already observed interval.

Federal Information Processing Standards (FIPS): Standards for information processing that comprise a
geographically exhaustive five digit code system wherein areas such as State, counties, territories, and
metropolitan areas are uniquely identified. FIPS codes were officially replaced with the Geographic
Names Information System (GNIS) as the Federal and national standard for geographic nomenclature.
However, FIPS codes are still maintained by the Census Bureau and are sometimes referred to as either
Census codes of Federal codes.

Final Payment: The last payment to a claimant which exhausts the individual's maximum potential benefit
entitlement under a specific program is referred to as a final payment.

Forgiveness Level: See Earnings Disregarded.

Gain: A weighting factor used in the Kalman Filter in determining the current month coefficients. Using
this factor, a portion of the difference between the current month's CPS and the preliminary model
estimate is added or subtracted from the previous month's coefficient value. This is used to produce the
current month's coefficient.

Geographic Names Information System (GNIS): The Geographic Names Information System (GNIS) is
the Federal and national standard for geographic nomenclature. The GNIS contains information about
physical and cultural geographic features of all types in the United States, associated areas, and
Antarctica, current and historical, but not including roads and highways. The GNIS Feature ID has
superseded the Federal Information Processing Standard (FIPS) 55 Place Code as the Federal feature
identifier.

Handbook Method: A building-block estimation method that uses data from several sources—including
the Current Population Survey, the Current Employment Statistics program, and unemployment insurance
program—to produce labor force estimates at the substate level. Estimates for counties and MCDs are
produced using this methodology.

Henderson Trend Filter (H13): A filtering procedure, based on moving averages, to remove the irregular
fluctuations from the seasonally-adjusted series, leaving the trend. It is part of a set of trend filters
developed by Robert Henderson (1916) for use in actuarial work that are used extensively by seasonal
adjustment packages such as X-12-ARIMA.

Household: As defined by the Bureau of the Census, a household is all persons who occupy a housing
unit. A housing unit is a room or group of rooms intended for occupancy as separate living quarters and
consists of either a separate entrance or complete cooking facilities for the exclusive use of the occupants.

ICON (Interstate Connection): A centralized computerized system of reporting and exchanging
unemployment insurance claims information between States.

Independent Variables: Variables used in the regression equation to predict the dependent variable, "Y".
The independent variables are usually termed the "X" variables.

Information Technology Support Center (ITSC): Established by the Department of Labor to assist all
state Unemployment Insurance agencies in the area of Unemployment Insurance Information Technology.

Initial Claim: Any notice of unemployment filed by an individual to initiate (1) a determination of
entitlement to and eligibility for compensation (a new claim), (2) a subsequent period of unemployment
within a benefit year or period of eligibility (an additional claim), or (3) a new claim filed to request a
determination of eligibility and establishment of a new benefit year within an existing spell of
unemployment (transitional claim).

Institutional Population: Persons residing in CPS-defined institutions, such as prisons, nursing homes,
juvenile detention facilities, or residential mental hospitals. Persons residing outside of these institutions
constitute the non-institutional population.

Insured Unemployment: Unemployment during a week for which waiting period credit or benefits are
claimed under the regular unemployment insurance compensation programs, supplemental extended
benefit programs, or the railroad unemployment insurance program, is considered insured.

Insured Unemployment Rate (IUR): The rate computed by dividing Insured Unemployed for the current
quarter by Covered Employment for the first four of the last six completed quarters.

Intercept: The value of "Y" (dependent variable) where the regression line crosses the "Y" axis. The
intercept is usually denoted by .

Interpolate: A method to estimate values of a function between two known values.

Interstate claim: A claim filed in one (agent) State based on monetary entitlement to compensation in
another (liable) State. The agent State is usually the claimant’s State of residence. The liable State is the
location of the establishment in which wage credits were earned.

Intrastate Claim: A claim filed in the same State in which the individual's wage credits were earned. A
nonresident of the State, filing an intrastate claim is called a commuter claimant.

Job Leavers: Individuals who quit or otherwise terminate their employment voluntarily and immediately
begin looking for work.

Job Losers: Unemployed persons who involuntarily lost their last job or who had completed a temporary job. This
includes persons who were on temporary layoff expecting to return to work, as well as persons not on temporary
layoff. Those not on temporary layoff include permanent job losers and persons whose temporary jobs had ended.

Kalman Filter: A statistical technique used in the signal-plus-noise models to adjust the model coefficient.
The coefficients are updated each month with new information using the Kalman Filter Technique. This
technique combines information from the model and CPS when making the new model estimate by taking
into account the relative accuracy of each.
Labor Force: The total of all civilians classified as employed and unemployed. The labor force, in
addition, includes members of the armed forces stationed in the United States.

Labor Market Area (LMA): An economically integrated area within which individuals can reside and find
employment within a reasonable distance or can readily change jobs without changing their place of
residence. LMAs include both the metropolitan and micropolitan areas delineated by the Office of
Management and Budget (OMB) and the small labor market areas defined by the Bureau of Labor
Statistics. The standards used by OMB to delineate metropolitan and micropolitan areas are reviewed and
updated after each decennial census, and the application of the updated standards to population data from
that census and commutation data from the American Community Survey (ACS) 5-year estimates ending
in the decennial year results in a comprehensive revision of OMB areas. Five years later, OMB areas are
further revised, based on the application of the same standards to the same decennial population data but
updated commutation data from the non-overlapping ACS 5-year estimates (i.e., those ending in the fifth
year of the decade). For small labor market areas, BLS examines commuting data for counties (cities and
towns in New England) not included within metropolitan or micropolitan areas shortly after OMB
announces its revisions, tying adjacent counties into multi-county configurations based on the same
thresholds used by OMB for its metropolitan and micropolitan areas

Labor Surplus Area: Defined under the Defense Manpower Policy No. 4A as an area with at least 120
percent of the national unemployment rate. (There is a variable floor and ceiling rate of 6% and 10 %.)

LAUS Estimate: The official BLS-published employment and unemployment estimates. For States, they
are based on the signal-plus-noise models. For areas, they are developed using the Handbook procedures
and are controlled to the State levels.

LAUS Redesign: A multi-year, multi-project initiative implemented with January 2015 estimates that
improved labor force estimates for State and substate areas. The redesign included improved time-series
models, model-based benchmarking, improved treatment of outliers, improvements to the smoothed
seasonal adjustment process, replacement of decennial census inputs with ACS estimates, updated
procedures for developing substate estimates, and the implementation of 2010-Census based
configurations for substate areas.

Least Squares: A basic regression n technique used to "fit" (calculate) a model equation to a time series of
data. There are several different types of least square calculations but all are based on minimizing the sum
of the squared differences between the data points and a regression line.

Liable/Agent Data Transfer (LADT): The record format used for the exchange of statistical data via the
Interstate Connection (ICON). The LADT record format was developed to accommodate the exchange of
data pertaining to: a) interstate weeks claimed; b) intrastate commuter weeks claimed; c) interstate initial
claims (new, additional and transitional); and, d) interstate reopened claims and claim transfers by ICON
for the exchange of interstate UI claims data among States.

Liable State: Any State against which a worker files a claim for compensation through the facilities of
another (agent) State is the liable State. The State location of the establishment in which wage credits are
earned is the liable State.

Link Relative Technique: A method for employment estimation that involves, for each estimating cell,
comparing the ratio of all employees in one month to all employees in the preceding month. The all
employee estimate for each month is obtained by multiplying this ratio by the all employee estimate for
the previous month. The technique is used in the CES estimating methodology.

Local Area Unemployment Statistics (LAUS): The Federal/State cooperative program under which
employment and unemployment estimates for States and local areas are developed. These estimates are
prepared by State Employment Security Agencies in accordance with BLS definitions and procedures.
They are used for planning and budgetary purposes, as an indication of need for employment and training
programs, and to allocate Federal funds under WIA, FEMA, etc.

Mass Layoff Statistics (MLS): A former Federal-State cooperative program that used a standardized,
automated approach to identify, describe, and track the effects of major job cutbacks, using data from
each State's unemployment insurance database. This program was discontinued in 2013.

Mass Layoff Event: A layoff in which 50 initial claims or more have been filed against an establishment
during a five-week period, regardless of the duration.

Mean Square Error (MSE): A measure of the total error that can arise in an estimate. It is equal to the
variance plus the bias squared. Mean square error is a more comprehensive measure of estimation error
than variance and is an important statistical and analytical tool.

Metropolitan Division (MD): A county or group of counties within a CBSA that contains a core with a
population of at least 2.5 million. A Metropolitan Division consists of one or more main/secondary
counties that represent an employment center or centers, plus adjacent counties associated with the main
county or counties through commuting ties.

Metropolitan Statistical Area (MSA): A CBSA associated with at least one urbanized area that has a
population of at least 50,000. The Metropolitan Statistical Area comprises the central county or counties
containing the core, plus adjacent outlying counties having a high degree of social and economic
integration with the central county as measured through commuting.

Micropolitan Area (MC): A CBSA associated with at least one urban cluster that has a population of at
least 10,000, but less than 50,000. The Micropolitan Statistical Area comprises the central county or
counties containing the core, plus adjacent outlying counties having a high degree of social and economic
integration with the central county as measured through commuting.

Migration: The permanent movement of an individual's residence from one location to another.

Model: A mathematical equation that describes how one or more random variables are related to other
(non-random) variables. In a time series, this relationship is computed over time. The LAUS signal-plusnoise models relate State CPS labor force estimates to different independent variables that show strong
correlations to the monthly estimates.

Monetary Determination: A written notice issued to inform an individual whether or not the individual
meets the employment and wage requirements necessary to establish entitlement to compensation under a
specific unemployment insurance program. If an individual is entitled, the weekly and maximum benefit
amounts the individual may receive are also determined.

Minor civil divisions (MCDs): The primary governmental or administrative divisions of a county in many
states (parishes in Louisiana) and the county equivalents in Puerto Rico. The MCDs in 12 states
(Connecticut, Maine, Massachusetts, Michigan, Minnesota, New Hampshire, New Jersey, New York,
Pennsylvania, Rhode Island, Vermont, and Wisconsin) also serve as general-purpose local governments
that can perform the same governmental functions as incorporated places. For the six New England States

(Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont) MCDs are the
geographic unit for developing Handbook estimates.

Moving Average: A continuous process that uses a series of calculations made by initially taking the
simple average, or arithmetic mean, of a consecutive number of items, and then dropping the first item
and adding the next item in sequence and averaging, so that the number of items in the series remains
constant.

New Claim: The first initial claim filed in person, by mail, telephone, or the Internet to request a
determination of entitlement to and eligibility for compensation. This is one of three types of initial
claims.

New England City and Town Area (NECTA): A statistical geographic entity used in New England that is
defined using cities and towns as building blocks and is conceptually similar to the Core Based Statistical
Areas (which are defined using counties as building blocks).

New England City and Town Area Division: A city or town or group of cities and towns within a NECTA
that contains a core with a population of at least 2.5 million. A NECTA Division consists of a main city
or town that represents an employment center, plus adjacent cities and towns associated with the main
city or town, or with other cities and towns that are in turn associated with the main city or town, through
commuting ties. Conceptually similar to Metropolitan Divisions.

New Entrants Unemployed: Individuals who enter the labor market for the first time and do not find jobs.
They include students entering the labor market after graduation from school and others who have not
previously held a full- time job lasting two weeks or longer.

Nonagricultural Wage and Salary Employment: In the CES program, this is a count of jobs by place of
work on nonagricultural establishment payrolls (including employees on paid sick leave, paid holiday, or
paid vacation) for any part of the pay period including the 12th of the month. It does not include
proprietors, self-employed, unpaid volunteer or family workers, domestic workers in households, military
personnel, and persons who are laid off, on leave without pay, or on strike for the entire reference period.

Noninstitutional Population: See Institutional Population.

Nonmonetary Determination: A process that determines whether a claimant meets legal criteria other than
wage credits under State UI law. It is usually concerned with: (1) reason claimant left job (separation
issues); and (2) job search (able, available, and actively seeking work).

Not in the Labor Force: All persons 16 years of age or older in the civilian noninstitutional population
who are neither employed nor unemployed are considered not in the labor force. Some examples are
students, housewives, retirees, etc.

Place-of-Residence Adjustment of Employment: Establishment-based data, which are on a place-of-work
basis, are adjusted to reflect the place of residence of the employed. The current adjustment also corrects
for multiple jobholding in the place-of-work series. See Dynamic Residency Ratio.

Population-Based Indexed Share Employment Disaggregation: A method that uses the annually prepared
total population estimates and data from the Census to disaggregate labor market area total employment
to the county or city level. This method is used only in conjunction with the claims-based unemployment
disaggregation.

Population Controls: Refers to population data developed from various independent sources, such as vital
statistics on births, deaths, migration, school enrollment, persons living in group quarters, inmates in
institutions, etc., which are used in Current Population Survey estimation procedures to independently
adjust sample-based labor force levels. Population controls are updated annually by the Bureau of the
Census and provided to the Bureau of Labor Statistics.

Population Estimates: Annual population estimates prepared by the Census Bureau that entails updating
population information from the most recent census with information found in the annual administrative
records such as tax records, Medicare records and some vital statistics information.

Predicted Value: The value of Y' (Y prime) that one obtains by “plugging in" values of the independent
variables into the formula for the regression line is the predicted value. The coefficients have already been
determined by a mathematical formulation.

Prediction Period: A period of time which is outside the sample period. Coefficients for the regression
line derived from the sample period are used to make predictions in subsequent periods. It is also called
the "outside sample" period.

Primary Sampling Unit (PSU): The first stage of CPS sampling involves dividing the United States into
primary sampling units, most of which comprise a metropolitan area, a large county, or a group of smaller
counties with homogeneous demographic and economic characteristics.

Program for Measuring Insured Unemployment Statistics (PROMIS): A stand-alone PC-based system
that stores all claimant information, including socioeconomic characteristics, and generates the UI inputs
to LAUS and, potentially, other programs. PROMIS operates as the clearinghouse for multi-purpose input
data, allowing flexibility to provide a more complete picture of the unemployment situation at substate
levels.

Quarterly Census of Employment and Wages Program (QCEW): A federal/State cooperative program that
produces a comprehensive tabulation of employment and wage information for workers covered by State
unemployment insurance (UI) laws and Federal workers covered by the Unemployment Compensation
for Federal Employees (UCFE) program.

Railroad Retirement Board (RRB): The RRB is an independent agency in the executive branch of the U.S.
government which administers a comprehensive social insurance system for the nation's railroad workers
and their families, providing protection against the loss of income resulting from old age, disability,
death, unemployment, and temporary sickness.

Raking: This is the process which forces additivity among components to the aggregate estimate. It is
performed on an iterative basis in the CPS.

Real-Time Benchmarking: A tiered approach to estimation in which the census division estimates are
benchmarked to the national levels of employment and unemployment on a monthly basis. The
benchmarked division model estimates are then used as the benchmark for the States within each division.
The distribution of the monthly benchmark adjustment to the States is based on each State's monthly
model estimate. In this manner, the monthly State employment and unemployment estimates will add to
the national level. Substate estimates are then revised and forced to add to the new State estimates. In the
past, this was done annually because the state data were benchmarked to the CPS annual average for each
state. Under this approach, benchmarking occurs monthly.

Reentrants In the CPS: Persons who previously worked at a full-time job at least two weeks but who were
out of the labor force for two weeks or more prior to beginning to look for work.

Reference Week: The week for which data are collected. For the CPS, the reference week is is defined as
the 7-day period, Sunday through Saturday that includes the 12th of the month. (On occasion, the
reference week in November and December may be week including the 5th of the month, to facilitate data
collection during the holiday period.)
Regression: A statistical tool which utilizes the relation between two or more variables so that one
variable can be predicted from the other(s).

Regression Equation: A statistical process for estimating the relationship between variables. In this
example, the equation has an intercept (), independent variables (X1 and X2) with coefficients (1 and
2 respectively and an error term,). The equation is YXX
Regression Line: A line fitted to the points in the scatter plot to summarize the relationship between the
variables being studied. When it slopes down (from top left to bottom right), this indicates a negative or
inverse relationship between the variables; when it slopes up (from bottom right to top left), a positive or
direct relationship is indicated.

Reopened claim: A claim filed after a break in claimed weeks during a benefit year. This break could be
caused by illness, disqualification, unavailability, or failure to report for any reason other than job
attachment. It is not a break resulting from other employment.

Residency Adjustment System (RAS): A National Office software system that assists States in correctly
coding the residency of their UI claims records. The system verifies and corrects erroneous addresses and
assigns geocodes including State, county, city/town, longitude and latitude and census tract and block.
RAS facilitates the city claims disaggregation process and is required for the use of PROMIS.

Rotation Group: One of eight systematic subsamples which comprise the total CPS sample. A rotation
group is in the sample for four consecutive months one year, leaves the sample during the following eight
months, and then returns for the same four calendar months of the next year.
Sample: A subset of a statistical population usually selected randomly for the purpose of making
generalized statements about the whole.

Sample Period: A period of time which is used to derive coefficients for the regression line. It is also
called the "inside sample" period.

Sampling Error: The measure of sampling variability, that is, the natural variations that might occur by
chance because only a sample of the population is surveyed.

Sample Regression: A type of regression in which the dependent variable is calculated from a sample
survey. Consequently there is an additional error (sampling error) to be considered.

Sampling Ratio: The proportion of units needed to be sampled to provide data of a specified level of
statistical reliability. Sampling ratios vary by cell, depending on the degree of variability of the measured
item.

Scatter Plot: A graph which plots the values of the dependent variable (Y) against the values of one of the
independent variable (X). By convention, the "X" variable is plotted against the horizontal scale and the
"Y" variable is plotted against the vertical scale.

Signal Extraction in ARIMA (Auto Regressive Integrated Moving Average) Time Series (SEATS): A
model based approach to seasonal adjustment that provides error measure. SEATS is used to adjust
metropolitan areas and divisions.

Self-Employment Assistance (SEA): An optional program to help unemployed workers to create their
own jobs by starting small businesses. To be eligible for the program an individual must be eligible for
unemployment compensation, have been permanently laid off from his/her previous job and identified
through the profiling system as likely to exhaust his/her benefits, and must participate in self-employment
activities including entrepreneurial training and business counseling.

Seasonal Adjustment: A statistical technique that eliminates the influences of weather, holidays, the
opening and closing of schools, and other recurring seasonal events from economic time series. This
permits easier observation and analysis of cyclical, trend, and other nonseasonal movements in the data.

Separation Issue: Nonmonetary determination in which the claimant acted in the termination of the
employment relationship. For example, voluntary quit without good cause, or voluntary quit for personal
reasons.

Series Break: An interruption in a time series caused either by a change in definition or in methodology
which makes it improper to compare data from after the change with data from before the change.

Short-time compensation: A program, commonly known as work-sharing, that allows an employer, faced
with the need for layoffs because of reduced workload, to reduce the number of regularly scheduled hours
of work for all employees rather than incur layoffs. This program provides partial UI benefits to
individuals whose work hours are reduced from full- time to part-time on the same job.

Signal-Plus-Noise Models: Econometric models used by the LAUS program to produce State labor force
statistics. The models measure the true labor force value contained in the monthly CPS estimates (the
signal) by extracting the noise associated with CPS sampling error.

Slope: A value that tells how much change in the dependent variable (Y) results from a change in one of
the independent variables (X). It is defined as the change in "Y" divided by the change in "X".

Smoothed Seasonal Adjustment (SSA): Seasonally-adjusted estimates that incorporate a long-run trend
smoothing procedure. A smoothed-seasonally adjusted series was introduced in 2010 to reduce the
number of spurious turning points in the former estimates. The estimates are smoothed using the
Henderson Trend Filter (H13) that suppresses irregular variation in real time.

Smoothing: In the time series regression, one month's data are used in estimating another and the best
estimate is made when data from all the other months are incorporated. The process of forward-backforward model re-estimation is referred to as smoothing because of its impact on monthly estimates. In
LAUS, smoothing is part of the annual benchmarking processing to update the model estimates series.

Standard Deviation: A measure of dispersion around the mean value of a population frequently denoted
by sigma (σ). It is the positive square root of the variance.

Standard Error: The term "Standard Error" can be used in many contexts. In general, it refers to the
variability of an estimate. In sampling, it usually refers to the confidence interval of the sample estimate the probability of including the true value with repeated sample. One standard error is about 68 percent
confidence; and 1.645 times the standard error is the more commonly used 90 percent confidence. The
model estimates also have confidence intervals. These relate to the variability of the estimate relative to
the regression line.

State Additional Benefits (AB): State financed programs for extending the potential duration of benefits
during periods of high unemployment for claimants in approved training who exhaust benefits, or for a
variety of other reasons. Although some state laws call these programs “extended benefits,” the term
“additional benefits” is used to avoid confusion with the federal-state EB program.

State Employment Security Agency (SESA): A generic name for the State agency usually responsible for
the following three activities: (1) The Unemployment Insurance Program which includes UI tax
collection, administration, and determination and payment of unemployment benefits. (2) The
Employment or Job Service Program which is an exchange for workers so seeking work and employers
seeking workers. (3) Research and Analysis which includes collection, analysis, and publication of labor
market information.

Statistical Population: A group of entities or individuals that are of concern to a statistician for a particular
investigation. This is sometimes referred to as simply a "population".

Stochastic: A term used to denote the randomness of a variable or process. A stochastic, or random,
variable is one whose value changes. In the case of the LAUS regression models, the values of the model
variables change from month to month.

Survey: The process used to collect data for the analysis of some aspect of a group or area.

Time Series: A consecutive set of observations over a specified period of time.

Time Series Independence: A condition present when successive values of a time series are nonrelated or
noncorrelated.

Trade Readjustment Allowances (TRA): Benefits provided to individuals who were laid off or had hours
reduced because their job was adversely affected by increased imports from other countries.

Transitional Claim: A new claim filed to request a determination of eligibility and establishment of a new
benefit year within an existing spell of unemployment. This is one of three types of initial claims.

Trend-Cycle Cascade Filter (TCCF): A combination of the Henderson filter and a seasonal filter that
suppresses the variability due to real-time benchmarking while simultaneously removing any residual
seasonality that may be present in the series

Unemployment Compensation for Ex-Servicemen (UCX): This federal program provides unemployment
benefits to ex-servicemen.

Unemployed: In the CPS, those individuals considered unemployed must be 16 years of age or older who
do not have a job but are available for work and are actively seeking work during the 4-week period
ending with the reference week (the week including the 12th of the month). The only exceptions to these
criteria are individuals who are waiting to be recalled from a layoff and individuals waiting to report to a
new job within 30 days. They are also considered unemployed.

Unemployed Disqualified: Persons who are able to work and are available for work but are disqualified
from receiving benefits for separation issues or other nonmonetary reasons.

Unemployment Compensation for Federal Employees (UCFE): This federal program provides benefits to
federal employees.

Unemployment Insurance (UI): Insurance premiums collected by the State and Federal governments from
which unemployment compensation is paid.

Unemployment Rate: The number of persons unemployed, expressed as a percentage of the civilian labor
force.

Variable: An entity that can take on a number of different values. It is frequently denoted by letters such
as "X" or "Y". Examples of variables would be CPS unemployment or CES employment.

Variable Coefficient Model (VCM): A type of sample regression model in which the model's coefficients
are allowed to change over time.

Variance: A mathematical measure of the dispersion of the values of a variable around its mean. The
variance may arise from a sampling of the population under study, or may just measure the variability of
population values around its means. The variance is frequently denoted as sigma squared (σ2).

Waiting Week: A period of unemployment during which a claimant may not draw benefits and during
which certain requirements essential to the establishment of claimant eligibility for benefits must be met.

Weeks Claimed: The number of weeks of benefits claimed, including weeks for which a waiting period or
fixed disqualification period is being served. Interstate claims are counted by State of residence.

Worksharing: See short-time compensation.

X-11 ARIMA and X-12 ARIMA: A nonparametric approach to seasonal adjustment. Used for seasonally
adjusting the national CPS data as well as LAUS estimates for some metropolitan areas. In 2003, BLS
adopted the use of X-12-ARIMA as the official seasonal adjustment procedure for CPS labor force series,
replacing the X-11-ARIMA program that had been used since 1980. Both X-12- and X-11-ARIMA are
based on earlier versions of the widely used X-11 method developed at the U.S. Census Bureau in the
1960s.