Local Area Unemployment Statistics Program

Local Area Unemployment Statistics
Program Manual

..and the model should fit CPS more closely.
Rate
10.0
CPS

MODEL

9.0
8.0

7.0
6.0
5.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

U.S. Department of Labor
Bureau of Labor Statistics
March 13, 2003

Table of Contents

LAUS Prorgam Manual

Table of Contents
Local Area Unemployment Statistics Program: Introduction 1-1
History 1-3
Data Sources 1-8
Summary of Estimation Methods 1-10
Uses and Publication of LAUS Estimates 1-12
Inputs to LAUS Estimation: The Current Population Survey 2-1
Introduction to the CPS 2-1
CPS Labor Force Concepts and Definitions 2-4
Reliability of CPS Estimates 2-8
Sample Design 2-12
Data Collection 2-19
Estimation Procedures 2-22
Inputs to LAUS Estimation: The Unemployment Insurance System 3-1
State Role 3-3
Differences: UI Data versus the CPS 3-5
BLS Standards for UI Data 3-6
Inputs to LAUS Estimation: Establishment Data Sources 4-1
The Current Employment Statistics Program 4-1
The Covered Employment and Wages Program 4-5
Differences: Establishment Data Sources versus the CPS 4-10
Uses of CES Data in the LAUS Program 4-11
Inputs to LAUS Estimation: Census Data 5-1
The Decennial Census: Enumerated and Sample-Based Data 5-1
Differences: Census versus CPS/LAUS Estimates 5-4
Uses of Decennial Census Data in LAUS 5-6

LAUS Program Manual

Table of Contents

Development of Statewide Estimates 6-1
Background 6-1
Signal-Plus-Noise Estimation Model 6-3
Description of the Employment Model 6-10
Description of the Unemployment Rate Model 6-12
Detailed Description of the Estimation Process 6-14
Seasonal Adjustment of Statewide Estimates 6-19
LAUS Estimation: Labor Market Area Estimates 7-1
Introduction 7-1
Labor Market Area Employment 7-6
Labor Market Area Unemployment 7-14
Monthly Step-3 Ratios 7-23
CPS Monthly Agricultural Factors 7-25
U.S. Survival Rates by Age 7-36
Seasonal Monthly A’ Factors 7-38
Seasonal Monthly B’ Factors 7-39
Quarterly Exhaustee Survival Rates 7-40
LAUS Estimation: Additivity 8-1
Adjustment to Independent Statewide Estimates—The Handbook Share
Method 8-3
Interstate Areas 8-5
LAUS Estimation: Disaggregation 9-1
Introduction 9-1
The Population-Claims Method of Disaggregation 9-3
Use of 1990 Census Data in Disaggregating Labor Force Estimates—
Census-Share Method 9-17
Annual Processing 10-1
Annual Model Review 10-1
Population Controls 10-3
Annual Re-Estimation or “Smoothing” 10-7
Benchmarking State Estimates 10-9
Benchmarking Sub-State Area Estimates 10-11

LAUS Program Manual

Table of Contents

STARS Macro Output 11-1
Introduction 11-1
STARS Cover Sheet 11-2
STARS Table 1: Year-to-Date Model Estimates 11-4
STARS Table 2: Changes from Prior Month/Year 11-5
Table 3: Components of Change 11-10
Table 4: Components of the Signal 11-12
Table 5: Input Data 11-13
Figures 1 and 2: Plots of Employment Level and Unemployment
Rate 11-14
STARS User’s Guide at SunGard 12-1
Introduction 12-1
Main Options 12-3
STARS Estimation Module 12-4
STARS Review Module 12-10
STARS Transmit Module 12-19
STARS Annual Processing Module 12-21

Glossary G-1
Index I-1

LAUS Program Manual

Table of Contents

LAUS Program Manual

LAUS

Program Overview

1

Local Area Unemployment
Statistics Program:
Introduction and Overview

he Local Area Unemployment Statistics (LAUS) program is a
Federal/State cooperative program which produces monthly
employment and unemployment estimates for approximately 6,700
geographic areas. The areas include all States, the District of Columbia,
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). Annual average data for all States are derived
directly from the CPS. Monthly estimates for all States, the District of
Columbia, New York City, Los Angeles-Long Beach, and the respective
balances of New York and California, are produced using estimating
equations based on time series and regression techniques. These models
combine current and historical data from the CPS, the Current
Employment Statistics (CES) program, and State unemployment
insurance (UI) systems. 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 labor market areas (other

LAUS Program Manual 1-1

than New York and Los Angeles, but including Puerto Rico) are produced using a
standard methodology called the “Handbook” method. This method also uses
data from several sources, including the CPS, CES, State UI systems, and the
decennial census, to create estimates which are then adjusted to the State CPSbased measures of employment and unemployment. Below the labor market area
level, estimates are prepared for all counties and cities of 25,000 population, using
disaggregation techniques based on decennial and annual population estimates
and current UI statistics.

1-2 LAUS Program Manual

History

History
For nearly fifty years, 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) 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 Current Population Survey
(CPS) but without the high cost of a household survey. As early as 1961, the 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, 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, BLS was publishing CPS-based 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, 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

LAUS Program Manual 1-3

History

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 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 up 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 applicability of the reliability criterion for 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 the 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

1-4 LAUS Program Manual

History

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 replaced with
computer-generated tabulations. Through such improvements as the use of placeof-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.

LAUS Program Manual 1-5

History

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.
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 Current
Population Survey, 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 time series 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.

1-6 LAUS Program Manual

History

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

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

LAUS Program Manual 1-7

History

(Continued) LAUS Time Line
Year

Historical Developments Related To LAUS

1983

Second round of UI Database Survey conducted; Quality Assurance Program instituted

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

Second generation of LAUS models introduced; 1990 Census data incorporated into LAUS; new CPS questionnaire and data collection method
implemented

1996

Direct-use States adopt model based estimation method; Handbook
method streamlined to 13 steps

1-8 LAUS Program Manual

Data Sources

Data Sources
LAUS estimates are designed to reflect the labor force concepts embodied in the
Current Population Survey (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 Current Population Survey (CPS); the State Unemployment Insurance (UI)
systems; the Current Employment Statistics (CES) program; the Quarterly Report
of Employment, Wages, and Employer Contributions (ES-202); and the Decennial
Census. 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 Bureau of Labor Statistics. 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 50,000 occupied households and yield estimates of demographic,
social, and economic characteristics of the population.
The Bureau of Labor Statistics 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. The CPS statewide
annual averages are used as benchmarks for LAUS estimates for all States, the
District of Columbia, the Los Angeles Metropolitan Area, and New York City.
(See Chapter 2 for more details on the CPS.)

Unemployment Insurance Systems
Under the Unemployment Insurance 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.

LAUS Program Manual 1-9

Data Sources

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.)

Current Employment Statistics and ES-202
Both the CES and ES-202 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 ES-202 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 ES-202, 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 ES-202
programs.)

Decennial Census
The Decennial Census is a universe count of the national population conducted
each decade by the Bureau of the Census. Though primarily intended to apportion
seats to the U.S. House of Representatives and for determining legislative district
boundaries, the census also is a source of socioeconomic and demographic data in
great geographic detail.
The LAUS program methodology uses decennial census data for adjusting
establishment-based employment estimates to residency-based employment
estimates, for estimating certain employment and unemployment components in
the Handbook methodology, and disaggregating or apportioning labor market area
estimates to smaller areas. (See Chapter 5 for additional details on the decennial
census.)

1-10 LAUS Program Manual

Summary of Estimation Methods

Summary of Estimation Methods
Monthly estimates of employment and unemployment are prepared for
approximately 6,700 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

Monthly labor force estimates for all States, the District of Columbia, the Los
Angeles-Long Beach metropolitan area, 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.
Estimates are benchmarked annually to the annual average CPS Statewide
employment and unemployment estimates. (See Chapter 6.)
Labor Market Area Estimates

States are divided into Labor Market Areas (LMAs) which exhaust the geographic
area of the State. Independent estimates are produced for all LMAs (except for
New York City and Los Angeles, as noted above) 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. Handbook
estimates are adjusted for additivity to the LAUS Statewide estimates to create the
official LMA estimates. LAUS estimates for sub-LMA areas, such as individual
counties within multi-county LMAs and cities with populations over 25,000, are
derived by a disaggregation technique using population estimates and UI statistics,
or data from the decennial census. (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 1-11

Summary of Estimation Methods

LAUS Estimation Techniques

Area

Estimation Method

50 States

Signal-plus-noise regression model

District of Columbia

Signal-plus-noise regression model

New York City, Balance of NY State

Signal-plus-noise regression model

Los Angeles, Balance of California

Signal-plus-noise regression model

Labor Market Areas (LMAs)

Handbook, Additivity

Sub-LMA Areas

Disaggregation

1-12 LAUS Program Manual

Uses and Publications of LAUS Estimates

Uses and Publication of LAUS Estimates
The uses and geographic detail of LAUS estimates are subject to changing
legislative requirements. Responsibility for methodology under the Federal-State
cooperative arrangement and requirements for publication have been provided to
BLS under OMB Statistical Policy Directive No. 11, “Standard Data Source for
Statistical Estimates of Labor Force and Unemployment.” The complete text of
this Directive is provided at the end of this Chapter.

Legislative Uses of LAUS Estimates.
The following tabulation, “Administrative Uses of Local
Area Unemployment Statistics”, presents information on
the programs which 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 $43,875.1 million
in Fiscal Year 2002. These programs are described in
greater detail in the section following the tabulation.
ADMINISTRATIVE USES OF LOCAL AREA UNEMPLOYMENT STATISTICS
FY 2002
Allocation Formulas/Qualifying
USER
Geographic Areas Used
Reference Period
Funding
Criteria
AGENCY/PROGRAM
(Millions)
DOL-ETA
Economically
$ 950.0
States and areas of
Most recent
State funding allocation for WIA Title
Disadvantaged Adults
substantial
program year (JulyII is based on the following
and Dislocated Workers
unemployment (ASUs).
June).
proportions: 1/3 on relative number
(Workforce Investment
An ASU is a contiguous
of unemployed in ASUs, 1/3 on
Act Title II--Adult
piece of geography,
relative excess number of unemployed
Education and Literacy)
consisting of counties,
(i.e., number of unemployed in excess
cities, and/or parts of
of 4.5 percent of labor force), and 1/3
each, with a population of
on relative number of economically
at least 10,000 and an
disadvantaged.
unemployment rate of at
least 6.5 percent. (7) (12)
Same as above for state funding, with
Youth Activities (Title I,
$ 1,128.0
States and ASUs. (7) (8)
Most recent
0.25% of funds allocated to “outlying
Chapter 4)
(9) (12)
program year (Julyareas.”
June).
$ 225.1
States and ASUs. (7) (8)
Most recent
Same as above, with 0.25% of funds
Youth Opportunity
Grants (Title I, Chapter
(9) (12)
program year (Julyallocated to “outlying areas.”
4)
June).
Dislocated Workers
$ 1,549.0
States and substate areas.
Most recent
State funding is based on the
(Title I, Chapter 5)
(7) (8) (9) (12)
program year (Julyfollowing proportions: 1/3 on relative
June) for
number of unemployed, 1/3 on
unemployed and
relative excess number of
excess unemployed; unemployed, and 1/3 on relative
most recent
number of unemployed for 15 weeks
calendar year for
or more. Also, 0.25% funds allocated
unemployed 15+
to “outlying areas.”
weeks.
Wagner-Peyser Act
$ 987.4
States. (10) (12)
Most recent
State funding algorithm is based on
(Title III, Subtitle A)
calendar year.
the following proportions: 2/3 on
relative number in labor force and 1/3
on relative number of unemployed.
Labor Surplus Areas
(1)
Counties, cities over
Most recent 2An area qualifies as a LSA when its
25,000 population, and
calendar year
average unemployment rate is 20
balance of counties. (12)
average.
percent or more above the national
rate (including Puerto Rico) for the
period, with the threshold being no
lower than 6 percent and no higher
than 10 percent.

LAUS Program Manual 1-13

Uses and Publications of LAUS Estimates

ADMINISTRATIVE USES OF LOCAL AREA UNEMPLOYMENT STATISTICS
USER
FY 2002
Allocation Formulas/Qualifying
Geographic Areas Used
Reference Period
AGENCY/PROGRAM
Funding
Criteria
(Millions)
DOL-ETA
Federal-State Extended
(2)
States. (7) (12)
Most recent 3
State is eligible to pay EB if: (1) the
Unemployment Benefits
months for total
seasonally adjusted total
(EB)
unemployment
unemployment rate (TUR) for the
trigger (TUR) or
most recent 3-month period is at least
most recent 13
6.5 percent, and at least 10 percent
weeks for insured
above the State TUR for the same 3unemployment
month period in either of the 2
trigger (IUR).
preceding years, or (2) the insured
unemployment rate (IUR) is at least 5
percent, and at least 120 percent of the
average IUR for the same 13-week
period in either of the 2 preceding
years.
FEMA
Emergency Food and
$ 140.0
Counties, cities, and
Most recent 12Jurisdictions qualify for FEMA
Shelter Program
balance of counties. (7)
month average.
funding if they meet one of the
(8) (12)
following criteria: (1) 18,000 or more
unemployed with a jobless rate of no
more than 1 percentage point below
the national rate, (2) 400-17,999
unemployed with a jobless rate of at
least 1.2 to 1.5 percentage points
above the national rate, or (3) 400 or
more unemployed with a poverty rate
of at least 11.7 percent.
Commerce-EDA
Public Works Program
$ 250.0
Areas defined by
Most recent 24An area qualifies if: (1) the
geographic/political
month average.
unemployment rate is at least one
boundaries, e.g., States,
percentage point above the national
cities, counties, Indian
rate, (2) the per capita income is 80
reservations. (7) (8) (9)
percent or less of the national average
(12)
per capita income, or (3) there is a
special need, as determined by EDA,
arising from actual or threatened
severe unemployment or economic
adjustment problems resulting from
severe short-term or long-term
changes in economic conditions.
Economic Adjustment
$ 41.0
Same geographic areas
Most recent 24Same qualifying criteria used in the
(Title 9)
used in the Public Works
month average.
Public Works Program
Program.
USDA
Temporary Emergency
$ 150.0
States. (7) (8) (9) (12)
Fiscal year average.
Farm commodities and funds are
Food Assistance
allocated based on the following
Program (TEFAP)
proportions: 3/5 on relative number of
persons in households below the
poverty line and 2/5 on relative
number of unemployed persons.
Welfare Reform Act-$ 21,170.0 States, metropolitan areas Generally 12-month Waivers are granted to areas with: (1)
Waivers to Food Stamp
(3)
(MAs), counties, cities
periods, but no less
an unemployment rate over 10 percent
Time Limits
Indian reservations, and
than 3 months for
for the latest 12-month (or 3-month)
specially designated areas unemployment rate.
period or (2) insufficient jobs.
(e.g., census tracts). (7)
Not specified for
(12)
insufficient jobs
criterion.
DOJ-INS
Immigration Act of
(4)
MAs and counties, cities
Most recent
Visas are granted for lower investment
1990
and subareas within MAs. calendar year or 12amounts in rural areas or areas with an
Employment Creation
month average.
unemployment rate at least 50 percent
Visas
above the national average.

LAUS Program Manual 1-14

Uses and Publications of LAUS Estimates

ADMINISTRATIVE USES OF LOCAL AREA UNEMPLOYMENT STATISTICS
USER
FY 2002
Allocation Formulas/Qualifying
Geographic Areas Used
Reference Period
AGENCY/PROGRAM
Funding
Criteria
(Millions)
DOD-DLA
Procurement Technical
$ 18.2
States, counties, cities,
Most recent 24An area qualifies for assistance if: (1)
Assistance (PTA)
and townships. (7) (12)
month average.
the unemployment rate is at least one
percentage point above the national
average for the most recent 24-month
period or (2) the per capita income is
80 percent or less of the State average.
HHS
Temporary Assistance to
(11)
States. (10)
Most recent 3States can access funds if they are
Needy Families
(5)
month average.
determined to be "needy," based on a
(TANF)—Contingency
seasonally adjusted unemployment
Fund Drawdown
rate that is at least 6.5 percent for the
3-month period and at least 110
percent of the rate for the
corresponding period in either of the 2
preceding calendar years; or if the
number of food stamp recipients
increases at least 10 percent during the
3-month period. TANF automatically
gives block grants (with an upper limit
of $71 million) to Puerto Rico.
TANF—Exemption
$ 17,128.0 States. (9) (10)
Not available.
In transitioning from welfare to work,
from Benefit Limitation
(11)
individuals are granted up to 6 weeks
(6)
for which a job search or participation
in a workfare program will be counted
as work. This time limit is extended
to 12 weeks if the State
unemployment rate is at least 50
percent above the national rate.
TANF automatically gives block
grants (with an upper limit of $71
million) to Puerto Rico.
Treasury
Riegle Community
$17.0
MAs, counties, cities, and Most recent 12An institution may qualify if (part or
Development and
possible sub-areas (e.g.,
month period before all of) its service area: (1) is located
Regulatory
census tracts). (7) (8) (9)
announcement of
within one unit of general local
Improvement Act of
(10) (12)
application period.
government, (2) has a contiguous
1994--Bank Enterprise
boundary, (3) (a) has a population of
Awards
4,000 or more, if in a metropolitan
area; (b) has a population of 1,000 or
more, if outside of a metropolitan
area; (c) is entirely within an Indian
reservation, (4) has a poverty rate of at
least 30 percent, and (5) has an
unemployment rate at least 1.5 times
the national rate. Puerto Rico is
treated like a State.
Riegle Community
$5.6
Same geographic areas
Same reference
Same qualifying criteria used for the
Development and
used for the Bank
period used for the
Bank Enterprise Award.
Regulatory
Enterprise Awards
Bank Enterprise
Improvement Act of
Awards.
1994—
Small and Emerging
CDFI Assistance
Component
North American
$ 6.0
Communities (discrete
Most recent 12Eligibility of CAIP financing
Development Bank
(11)
geographical areas) i.e.,
month average
includes: (1) a significant job loss
(NADBank) Community
counties, towns, or cities.
connected to the passage of NAFTA
Adjustment and
and (2) a substantial continued need
Investment Program
for transition assistance.
(CAIP)

LAUS Program Manual 1-15

Uses and Publications of LAUS Estimates

ADMINISTRATIVE USES OF LOCAL AREA UNEMPLOYMENT STATISTICS
USER
FY 2003
Allocation Formulas/Qualifying
Geographic Areas Used
Reference Period
AGENCY/PROGRAM
Funding
Criteria
(Millions)
ARC (Appalachian
Regional Commission)
Distressed County Non$14.0
All of West Virginia and
Most recent 3-year
An area qualifies if its: (1) per capita
Highway Program
parts of 12 other states,
period for which
income is 2/3 the national average or
(DCNHP)
by county.
data are available at
less, (2) poverty rate is at least 1.5
the beginning of
times the U.S. rate, and (3)
application process.
unemployment rate is at least 1.5
times the national average.
General Area
Same qualifying criteria used for the
$34.0
Same geographic areas
Same reference
DCNHP.
Development Program
used for the DCNHP.
period used for the
DCNHP.
Distressed City Initiative
$2.0
Same geographic areas
Same reference
Same qualifying criteria used for the
used for the DCNHP.
period used for the
DCNHP.
DCNHP.
Small Business
Administration
Historically
(1)
Census tracts, nonMost recent annual
An area qualifies if it is: (1) a
Underutilized Business
metropolitan counties, or
average for
"qualified" census tract (as defined in
Zones (HUBZones)
Indian reservations. (7)
unemployment rate.
the 1986 IRS code), (2) a non(12)
metropolitan county with (a) median
household income less than 80% of
the statewide non-metropolitan
median or (b) an unemployment rate
at least 140% of the statewide
average, or (3) within the boundaries
of an Indian reservation.
HUD
Youthbuild Program
$59.8
Census tracts, nonNot specified.
An area can qualify if it is an
metropolitan counties.
underserved area, which is defined as
an area comprised of census tracts
with the following distress criteria: (i)
a census tract where the
unemployment remains high (50
percent or more above the nation’s
unemployment rate) and (ii) a census
tract where high rates of poverty
persists.
Total Appropriations
$ 43,875.1
NOTE: The term “cities” also includes townships and boroughs in selected states for various programs.
(1) Program does not allocate funds, but gives preference to firms in bidding on federal procurement.
(2) Under regular state extended benefits, monies are not appropriated, but are drawn from the Unemployment Insurance Trust Fund.
If the 3-month average TUR is at least 8%, and at least 10% above the TUR for the same 3-month period in either of the 2 preceding
years, the State enters a "high unemployment period" during which 20 weeks of EB are payable.
(3) Dollar amount is full cost of Food Stamp Program. Soup Kitchen and Food Bank funding was merged into the Welfare Reform
Act of 1996, and, though the program may continue to receive donations, there is no separate funding.
(4) Under IMMACT, a total of 3,000 visas are distributed to eligible immigrant entrepreneurs who establish a new commercial
enterprise in a targeted employment area (rural area or other area with high unemployment)
(5) Under the Welfare Reform Act, a Contingency Fund of State Welfare Programs was established, with a $2 billion limit for FY
1997-2001.
(6) Dollar amount is the full cost of the TANF program.
(7) The District of Columbia and Puerto Rico are considered states.
(8) Outlying areas include the U.S. Virgin Islands, Guam, American Samoa, Northern Marianas Islands, Marshall Islands,
Micronesia, and Palau.
(9) Native American Program includes Indians, Native Hawaiians, and Alaska Natives.
(10) The District of Columbia is considered a state.
(11) Currently funded by previous grants and awaiting new legislation. Dollar amount shown pertains to FY 2001.
(12) Program treats Puerto Rico as a state, and its areas as substate areas.
July 10, 2002

LAUS Program Manual 1-16

Uses and Publications of LAUS estimates

LAUS Program Manual 1-17

Uses and Publication of LAUS Estimates

Department of Commerce:
Economic Development Administration,
Public Works

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 EDA-designated 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

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

1-18 LAUS Program Manual

Uses and Publication of LAUS Estimates

strategically targeted business development and financing programs including
grants for revolving loan funds, infrastructure improvements, organizational
development, and market or industry research and analysis.

Department of Labor:
Employment and Training
Administration, Employment Service

Program Objectives: The employment
service is available to all those legally
authorized to work in the United States in
order to assist millions of job seekers and
employers and, in some areas, provide job
training and related services. The Federal
Government, through the Employment and
Training Administration, provides general
direction, funding, and oversight, and also assists the States with programs of test
development, occupational analysis, and maintenance of an occupational
classification system. The State employment security agencies operate 1,800
local Employment Service offices. In accordance with their needs, States may
provide specialized assistance to such groups as youth ages 16-22, women, older
workers, persons with disabilities, rural residents and workers, and the
economically disadvantaged.
Public employment service assistance, including employability assessment and
referral to training if necessary, is free to job seekers. Most of the service’s
appropriations come from the trust funds collected under the Federal
Unemployment Tax Act (FUTA), with a small portion coming from general
revenues.
Employment and Training Administration, Dislocated Workers

Program Objectives: To assist dislocated workers to obtain unsubsidized
employment through training and related employment services using a
decentralized system of State and local programs. The Economic Dislocation and
Worker Adjustment Assistance Act (EDWAA) provides funds to States and local
substate grantees. The Act authorizes employment and training help for
dislocated workers. Workers who lose their jobs in mass layoffs or plant closings
and others who have been laid off and are unlikely to return to their jobs can take
advantage of early intervention programs, occupational skill training, job search
assistance, support services, and relocation assistance.

LAUS Program Manual 1-19

Uses and Publication of LAUS Estimates

Employment and Training Administration, Job Training Partnership Act,
Title II-A: Adult Training Program

Program Objectives: To provide job training and related assistance to
economically disadvantaged individuals and others who face significant
employment barriers. The ultimate goal of the Act is to move trainees into
permanent, self-sustaining employment. This legislation authorizes training and
services for the economically disadvantaged and others who face significant
employment barriers. Training is afforded through grants to States for local
training and employment programs. States are responsible for further allocating
funds to their Service Delivery Areas (SDAs) and for overseeing the planning and
operation of local programs. Program services include an assessment of an
unemployed individual’s needs and abilities and a strategy of services such as
classroom training, on-the-job training, job-search assistance, work experience,
counseling, basic skills training, and support services.
Employment and Training Administration, Job Training Partnership Act,
Title II-B: Summer Youth

Program Objectives: To provide a summer youth employment and training
program for economically disadvantaged youths. The purpose of the Summer
Youth Employment and Training Program is to enhance the basic educational
skills of youth, encourage school completion or enrollment in supplementary or
alternative school programs, provide eligible youth with exposure to the world of
work, and enhance the citizenship skills of youth. Programs offer these
individuals jobs and training during the summer. This includes work-experience
programs and support services such as transportation. Academic enrichment also
is a major part of the program and may include basic and remedial education.
Employment and Training Administration, Job Training Partnership Act,
Title II-C: Disadvantaged Youth

Program Objectives: To provide for the long term employability of youth;
enhance educational, occupational, and citizenship skills; encourage school
completion or enrollment in alternative school programs; reduce welfare
dependency; and assist in the transition from school to work, apprenticeship,
military, or post secondary education and training. Program services may include
all authorized adult services, limited internships in the private sector school-towork transition services, and alternative high school services.
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, except for

1-20 LAUS Program Manual

Uses and Publication of LAUS Estimates

two states-- Massachusetts and Washington --which pay up to 30 weeks of
benefits. 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, 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.

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 decisionmaking, 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.

LAUS Program Manual 1-21

Uses and Publication of LAUS Estimates

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 and requires the Secretary to use $100 million yearly from the Food
Stamp account to purchase commodities for TEFAP during Fiscal Years 1997
through 2002.
Food, Nutrition, and Consumer Services—Food Stamps

Program Objectives: The Personal Responsibility and Work Opportunity
Reconciliation Act of 1996 limits receipt of Food Stamp benefits to 3 months in a
3-year period for able-bodied 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.

1-22 LAUS Program Manual

Uses and Publication of LAUS Estimates

Department of Justice:
Immigration and Naturalization Service,
Immigration Act of 1990 (IMMACT)

Program Objectives: To make 10,000 visas
available each fiscal year to qualified
immigrants seeking to enter the U.S. for the
purpose of engaging in a new commercial
enterprise. The new commercial enterprise
may take any lawful business form and
must both benefit the U.S. economy and
create full-time employment for not fewer
than 10 U.S. citizens, lawful permanent
residents, or other immigrants lawfully authorized to be employed.
To encourage the establishment of new enterprises in areas which would most
benefit from employment creation, 3,000 of the employment creation visas are
reserved for qualified aliens who have made investments in “targeted employment
areas.” Such areas are defined to include rural areas and areas which have
experienced high unemployment. A rural area is defined as any area other than an
area within a metropolitan statistical area (MSA) or within the outer boundary of
any town having a population of 20,000 or more. An area of high unemployment
under the Act is defined as a non-rural area with an average unemployment rate of
150 percent of the national average in the previous calendar year. Alternatively, a
letter from an authorized body of the government may certify that the area has
been designated a high unemployment area.

Department of Defense:
Defense Logistics Agency—Office of
Small and Disadvantaged Business
Utilization

Program Objectives: To provide funding
assistance to civil jurisdictions and nonprofit agencies working with small and
disadvantaged businesses. The purpose of
the Procurement Technical Assistance
(PTA) Cooperative Agreement Program is
to (1) generate employment and improve
the general economy of a locality by
assisting business firms in obtaining and performing under Federal, State, and
local government contracts; (2) increase Department of Defense assistance for

LAUS Program Manual 1-23

Uses and Publication of LAUS Estimates

eligible entities furnishing PTA to business entities; and (3) assist eligible entities
in the payment of the costs of establishing and carrying out new PTA programs
and maintaining existing PTA programs.

Department of Health and
Human Services:
Administration for Children and
Families, Temporary Assistance for
Needy Families

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.
Office of Community Services, Community Food and Nutrition Program

Program Objectives: The Community Food and Nutrition Program provides
assistance to public and private agencies at the community, local, and national
levels to meet the nutrition needs of low-income people. This is done by
coordinating existing food assistance resources, assisting in identifying sponsors
of child nutrition programs, and initiating new programs in under-served and unserved areas and developing innovative approaches at the State and local levels.

1-24 LAUS Program Manual

Uses and Publication of LAUS Estimates

Department of the Treasury:
Community Development Financial
Institutions Fund, 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
North American Development Bank (NADBank), Community Adjustment
and Investment Program

Program Objectives: To finance community adjustment and investment efforts
throughout the United States and Mexico. U.S. appropriations of $225 million
will be leveraged into financing for community adjustment projects that will
provide significant benefits for U.S. citizens and businesses. The NADBank's U.S.
community adjustment window will operate nationwide to offer financing directly
through existing federal credit programs to assist communities and businesses
adjust to the new trade environment created by NAFTA.

Appalachian Regional Commission:
Non-Highway Program

Program Objectives: The Commission was established to assist in the long-term
development of the chronically depressed region. Its main objectives are the
creation of new jobs and preparation of the people in the region to compete for
jobs wherever they choose to work and live. The non-highway program focuses
on the creation of new jobs and private investments and special help for the
region’s poorest or distressed counties. New jobs and private investment are
encouraged by grants supporting education, water and sewer services for
industrial and commercial needs, housing, small business development, health
care, development of natural resources, and research on topics directly related to
the region’s economic development.

LAUS Program Manual 1-25

Uses and Publication of LAUS Estimates

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 other than
small.

Environmental Protection Agency:
EPA Acquisition Regulations [proposed rules]

Program Objectives: The Environmental Protection Agency is amending its
Acquisition Regulation to include a new regulation on Socioeconomic Contract
Clauses. These amendments will allow contractors to receive technical points in
the bidding process for their use of local employment and training while
performing under EPA contracts. The use of the solicitation provision and contract
clause is expected to aid in decreasing the local unemployment rate and support
economic development in the area where contractual requirements will be
performed.

1-26 LAUS Program Manual

Uses and Publication of LAUS Estimates

Publication of LAUS Estimates
Data from the LAUS program are made available
to users in a variety of ways. Labor force and
unemployment data are published monthly for all
States and selected metropolitan areas in a news
release entitled” State and Metropolitan Area
Employment and Unemployment” and in the
Bureau’s periodical Employment and Earnings.
Estimates of labor force, employment and
unemployment for all States, metropolitan areas,
labor market areas, counties, cities with a
population of 25,000 or more, and other areas used
in the administration of various Federal economic
assistance programs, are provided in “Unemployment in States and Local Areas”,
which is available monthly in microfiche form by subscription from the U.S.
Government Printing Office.
Annual average employment status data are
provided each year in a press release entitled “State
and Regional Unemployment, Annual Averages”,
which is typically issued in the spring. It presents
data on the population, civilian labor force,
employed, unemployed, and unemployment rate
for regions, divisions, and States. Annual average
information for States and selected metropolitan
areas is also published each spring in Employment
and Earnings.
The annual publication, Geographic Profile of
Employment and Unemployment, provides annual average data for census regions
and divisions, the 50 States and the District of Columbia, 50 large metropolitan
areas, and 17 central cities. Data are provided on the employed and unemployed
by selected demographic and economic characteristics.
Current and historical data from the LAUS program are also available on-line via
LABSTAT, the Bureau’s public database. (Access via anonymous FTP or Gopher
is stats.bls.gov; via World Wide Web, stats.bls.gov/blshome.html.)

LAUS Program Manual 1-27

Office of Management and Budget Statistical Policy Directive Number 11

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 Bureau of Labor
Statistics, 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 Bureau of Labor Statistics 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 Bureau of Labor Statistics, 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 Bureau of Labor Statistics 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.

1-28 LAUS Program Manual

Office of Management and Budget Statistical Policy Directive Number 11

3. Data Publication

The Bureau of Labor Statistics, 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 Bureau of Labor Statistics shall
contain data for the Nation as a whole, and for each State and each local area for
which the Bureau of Labor Statistics has agreed to publish data. No agreement
between the Bureau of Labor Statistics 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 Bureau of Labor Statistics. 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 rate.

LAUS Program Manual 1-29

Office of Management and Budget Statistical Policy Directive Number 11

b. Current Data. For the purposes of this Directive, the term current data means
the most current, complete data published by the Bureau of Labor Statistics.
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.

1-30 LAUS Program Manual

Current Population

Survey (CPS)

2

Inputs to LAUS Estimation:
The Current Population Survey

Introduction to the CPS
he Current Population Survey (CPS) is a monthly household survey
conducted by the Bureau of the Census for the Bureau of Labor Statistics
(BLS). The CPS collects current labor force status about the 16+ civilian
noninstitutional population of the United States. It is a cooperative BLS/Census
effort with its design and methodology jointly planned by both bureaus.
Responsibilities for the survey are divided; the Census Bureau conducts data
collection, the BLS analyzes and publishes the data. The CPS data are used
directly to produce demographic labor force estimates for the nation and indirectly
as input to the models used to develop labor force estimates for States.

T

For current official State labor force estimation, monthly CPS estimates are
included in the LAUS signal/noise estimation procedure. Annual average CPS
data for States and selected areas are used as annual benchmarks. In addition,
national monthly and annual average CPS data are inputs to area Handbook line
items. For the Handbook estimation of Labor Market Areas (LMAs), various
ratios for estimating components of employment and unemployment not available
at the area level are developed using monthly and annual average data.

Background
The CPS has its origin in a program set up in 1940 to provide direct measurement
of unemployment each month from a sample survey. Several earlier attempts to
estimate the number of unemployed used various devices, ranging from guesses to
enumerative counts. During the latter half of the 1930’s, the Work Projects
Administration (WPA) (the Works Progress Administration prior to 1939) began
developing techniques for measuring unemployment. The Enumerative Check
Census, taken as part of the 1937 unemployment registration, was the first to

LAUS Program Manual 2-1

Introduction to the CPS

estimate unemployment on a nationwide basis using probability sampling.
This research and the experience with the Enumerative Check Census led
to the Sample Survey of Unemployment which was started in March 1940
as a monthly activity by WPA. The survey was transferred to the Bureau
of the Census in 1942 and its title was changed to The Monthly Report on
the Labor Force. The survey was renamed as The Current Population
Survey in 1948. BLS assumed responsibility for publication and analysis
of CPS data in 1959.
The CPS is the oldest continuous household survey in the world.
Throughout its history, the CPS has constantly been improved and
updated to keep pace with statistical and technological advances. Major
changes that have occurred in the CPS include improved identification of
households covered in the sample, improvements to sample design and
methodology, improved estimation procedures, and modifications to the
questionnaire and interview process.
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 which is in use today.
Throughout the 1970’s, a series of State-only sample expansions were
undertaken, in response to greatly expanded data needs at the subnational
level. While providing reliability needed to introduce CPS annual
average benchmarks in all States, this method was recognized as an
inefficient way of developing State estimates. In 1985, the national-based
design was changed to a State-based sampling design. A requirement of
this design was that annual average State estimates fall within specified
levels of reliability, while maintaining the current reliability of the
national estimates.
In 1994, computer assisted telephone interviewing (CATI) and computer
assisted personal interviewing (CAPI), as well as a new questionnaire
design, were phased in as part of continuing survey improvement.

Survey Process
The CPS survey process consists of three main phases. These phases are
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.

2-2 LAUS Program Manual

Introduction to the CPS

During the data collection phase, households are asked about activities during the
week that contains the 12th day of the month, the reference week. A
questionnaire is completed, either by a personal interview or by phone, for each
household member 16 years of age and over to determine the labor force status for
the previous week.
The goal of the estimation process is to take sample data and make estimates for
the population as a whole. Estimation involves a number of steps, including data
editing and imputation, basic weighting, noninterview adjustment, ratio
adjustment, compositing of estimates, and seasonal adjustment.

LAUS Program Manual 2-3

CPS Labor Force Concepts and Definitions

CPS Labor Force Concepts and Definitions
The CPS classification of persons as “employed,” “unemployed,” or “not
in the labor force” are used in the LAUS estimation methodology. These
classifications are based on a person's labor force status during the survey
reference week (the week including the 12th of the month). The CPS
questionnaire is designed to first determine whether a person is employed.
If a person is not employed, the next series of questions are designed to
determine whether the person is unemployed and still in the labor force,
or out of the labor force entirely.
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
thus produces an unduplicated count of employed or unemployed persons.
The CES survey, on the other hand, is establishment-based and is
designed to produce counts of the number of jobs in the economy.
Therefore, persons holding more than one job could be counted more than
once depending on which establishments were in the survey sample.
Since the LAUS program uses employment numbers from both sources, a
reconciliation must be made to adjust CES data to a residency base. (See
Residency Adjustment Ratio.)

Labor Force
Civilian Noninstitutional Population (CNP): The CNP is the base
population used in the calculation of labor force statistics. This category
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 (e.g. penal
and mental facilities, homes for the aged) and who are not on active duty
in the Armed Forces.
Labor Force: The labor force is comprised of all persons classified as
employed or unemployed with respect to the criteria described below.
Not in the Labor Force: This category includes all persons in the civilian
noninstitutional population who are not classified as employed or
unemployed. This classification is based on information about their
desire and availability for work collected during the CPS interview, job
search activity in the prior year, and reason for not looking in the 4-week
period ending with the reference week. This group includes the
discouraged workers category which is described below.
• Discouraged workers are defined as persons not in the labor force who
want and are available for a job and who have looked for work

2-4 LAUS Program Manual

CPS Labor Force Concepts and Definitions

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.

Employed and Unemployed
Employed Persons: There are two categories of employed persons. The first
category includes all civilians who, during the survey week, did any work at all as
paid employees 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. The second category includes all those who were not
working but who had jobs or businesses from which they were temporarily absent
because of illness, bad weather, vacation, child care problems, maternity or
paternity leave, labor-management disputes, job training, or other family or
personal reasons, whether 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. Included in the total are employed
citizens of foreign countries who are residing in the United States, but are not
living on the premises of an embassy. Excluded are persons whose only activity
consisted of work around their own home (such as housework, painting, repairing,
etc.) or volunteer work for religious, charitable, or similar organizations.
Employed workers are also defined by the following class-of-worker groupings:
• Wage and salary workers who receive wages, salary, commissions, tips, or
pay-in-kind from an employer. This category is further subdivided into private
and government workers.
• Self-employed persons who work for profit or fees in their own business,
profession, or trade, or farm. Only the unincorporated self-employed are
included in this category. Self-employed persons who report that their
business is incorporated are included among the wage and salary workers
because they are technically paid employees of a corporation.
• Unpaid family workers are persons who work 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.
• Multiple jobholders are employed persons who, during the reference week, had
either two or more jobs as a wage and salary worker, were self-employed and
also held a wage and salary job, or worked as an unpaid family worker and
also held a wage and salary job. A person employed only in private
households (cleaner, gardener, baby-sitter, etc.) who worked for two or more

LAUS Program Manual 2-5

CPS Labor Force Concepts and Definitions

employers during the reference week is not counted as a multiple
jobholder, since working for several employers is considered an
inherent characteristic of private household work. Also excluded are
self-employed persons with multiple businesses and persons with
multiple jobs as unpaid family workers.
Unemployed Persons: This category includes all persons who (1) had no
employment during the reference week, (2) were available for work
except for temporary illness, and (3) had made specific efforts, such as
contacting employers, to find employment some time during the 4-week
period ending with the reference week.
Only active methods—which have the potential to result in a job offer
without further action on the part of the jobseeker—qualify as “efforts to
find employment.” Examples include going to an 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.
Examples of the “other” category 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.
Persons waiting to be recalled to a job from which they were laid off need
not be looking for work to be classified as unemployed.
Unemployed persons are further categorized according to their status at
the time they began their current job search. Five major reasons for
unemployment are defined as follows:
• Job losers, comprised of persons on temporary layoff, who have been
given a date to return to work or who expect to return within six
months (persons on layoff need not be looking for work to qualify as
unemployed), and permanent job losers, whose employment ended
involuntarily and immediately began looking for work.
• Job leavers, persons who quit or otherwise terminated their
employment voluntarily and are looking for work.
• Persons who completed temporary jobs, who began looking for work
after the jobs ended.
• Re-entrants, persons who previously worked but were out of the labor
force prior to beginning their job search.
• New entrants, persons who never worked.

2-6 LAUS Program Manual

CPS Labor Force Concepts and Definitions

All unemployed persons who made specific efforts to find a job sometime during
the 4-week period preceding the survey week are classified as job seekers. Job
seekers do not include those person classified as on temporary layoff, who
although often looking for work, are not required to do so to be classified as
unemployed.
Unemployment Rate: This rate represents the number unemployed as a percent of
the civilian labor force. This measure can also be computed for groups within the
labor force classified by sex, age, race, Hispanic origin, marital status, etc.
Duration of Unemployment: Duration represents the length of time through the
current reference week that persons classified as unemployed had been looking
for work, and thus is a measure of an in-progress spell of joblessness. For persons
on layoff, duration of unemployment represents the number of full weeks they had
been on layoff. Two useful measures of the duration of unemployment are the
mean and the median. Mean duration is the arithmetic average computed from
single weeks of unemployment. Median duration is the midpoint of a distribution
of weeks of unemployment.

LAUS Program Manual 2-7

Reliability of CPS Estimates

Reliability of CPS Estimates
There are two types of errors possible in an estimate based on a sample
survey—sampling and nonsampling. The mathematical discipline of
sampling theory provides methods for estimating standard errors when
the probability of selection of each member of a population can be
specified. The standard error, a measure of sampling variability, can be
used to compute confidence intervals that indicate a range of differences
from true population values that can be anticipated because only a sample
of the population has been surveyed. Nonsampling errors, such as
response variability, response bias, and other types of bias occur in
complete censuses as well as sample surveys. In some instances,
nonsampling error can be more tightly controlled in a well-conducted
survey where it is feasible to collect and process the data more skillfully.
Reinterview programs are often used to measure response variability and
response bias. Estimation of other types of bias is one of the most
difficult aspects of survey work, and often adequate measures of bias
cannot be made

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 nonsampling error has been found to be small on
estimates of change, such as month-to-month change. Estimates of
monthly levels are generally more severely affected by nonsampling error.
Response error, nonresponse error, error in independent population
controls, processing error, and coverage error are types of nonsampling
error that affect the CPS.
Response Error. CPS estimates are subject to response errors made
during data collection. These errors include 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. Errors
occurring during the interview phase of the survey are studied by means
of a reinterview program. This program is used to estimate various
sources of error as well as to evaluate and control the work of the
interviewer. A random sample of each interviewer’s work is inspected
through reinterview at regular intervals. The results indicate, among other
things, that the data published from the CPS are subject to moderate
systematic biases.

2-8 LAUS Program Manual

Reliability of CPS Estimates

Nonresponse Error. In a typical month, about 6-7 percent of occupied sample
households are not interviewed because residents are not at home, refuse to
cooperate, or are unavailable for other reasons. During estimation, 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 a 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.
Independent Population Controls. 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. Population controls used in CPS estimation are derived from the 1990
census counts, adjusted for the census undercount. Errors may arise in the
independent population estimates because of underenumeration of certain
population groups or errors in age reporting in the last census, or similar problems
in the components of population change (mortality, immigration, etc.) since that
date. Also, errors in estimated components of change since the last census affect
the accuracy of inter-censal population estimates; these in turn add to error in CPS
estimates. The Census Bureau used data from the 1990 Post Enumeration Survey
(a sample survey) to estimate the undercount. The adjusted census population
estimates are thus subject to sampling error as well as nonsampling error.
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. Undercoverage 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 (adjusted for the undercount). It is known that the
CPS undercoverage varies with age, sex, race, and Hispanic origin. Generally
undercoverage is larger for men than for women and larger for blacks, Hispanics,
and other races 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.

LAUS Program Manual 2-9

Reliability of CPS Estimates

Sampling Error
When a sample rather than the entire population is surveyed, estimates
differ from the true population values that they represent. This difference,
or sampling error, occurs by chance, and its variability is measured by the
standard error of the estimate. Sample estimates from a given survey
design are considered unbiased when an average of the estimates from all
possible samples would yield, hypothetically, the true population value.
In this case, the sample estimate and its standard error can be used to
construct approximate confidence intervals, or ranges of values, that
include the true population value with known probabilities.
If the process of selecting a sample from the population were repeated
many times and an estimate and its standard error calculated for each
sample, then:
1.) Approximately 68 percent of the intervals from one standard
error below the estimate to one standard error above the
estimate would include the true population value.
2.) Approximately 90 percent of the intervals from 1.6 standard
errors below the estimate to 1.6 standard errors above the
estimate would include the true population value.
3.) Approximately 95 percent of the intervals from two standard
errors below the estimate to two standard errors above the
estimate would include the true population value.

Although, as indicated above, the estimating methods used in the CPS do
not produce unbiased estimates, biases for most estimates are believed to
be small enough so that these confidence interval statements are
approximately true. Since it would be too costly to develop standard
errors for all CPS estimates, generalized variance function techniques are
used to calculate sets of standard errors for various types of labor force
characteristics. Standard errors computed from these methods reflect
contributions from sampling errors and some kinds of nonsampling errors
and indicate the general magnitude of an estimate’s standard error rather
than its precise value. Standard error tables are provided in the monthly
publication Employment and Earnings.

CPS Monthly and Annual Reliability Criterion
Data reliability is measured by calculating the coefficient of variation
(CV) of the unemployment level, where 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.

2-10 LAUS Program Manual

Reliability of CPS Estimates

The sample design maintains a 1.9 percent CV on national monthly estimates of
unemployment level. A 6-percent unemployment rate is assumed. This means a
month-to-month change in the unemployment rate must exceed 0.2 percentage
point 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 at most 8 percent 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 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 4 percent.

LAUS Program Manual 2-11

Sample Design

Sample Design
Introduction
Since the inception of the survey, there have been various changes in the
design of the CPS sample. The sample is redesigned and a new sample
selected after each decennial census. Also, the number of sample areas
and the number of sample persons are changed occasionally. Most of
these changes are made in order to improve the efficiency of the sample
design, increase the reliability of the sample estimates, or control cost.
The 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, sample areas, called Primary Sampling
Units (PSUs), are chosen. In the second stage, Ultimate Sampling Unit
(USUs) clusters, composed of about four housing units each, are selected.
Sample sizes and sampling rates are determined by the specified
reliability requirements. A sample rotation scheme is used to improve
reliability of month-to-month and year-to-year change estimates without
overburdening any specific group of households with an unduly long
period of inquiry.

Selection of PSUs (First Stage of Sampling)
The entire area of the United States, consisting of 3,141 counties and
independent cities, is divided into 2,007 primary sampling units (PSUs).
There are several criteria used for determining PSUs.
1.) PSUs are defined within States and do not cross State
boundaries.
2.) In most States, a PSU consists of a county or a number of
contiguous counties. In New England and Hawaii, minor civil
divisions are used instead of counties.
3.) Metropolitan areas within a State are used as a basis for forming
many PSUs.
4.) They are usually less than 3,000 square miles, with population of
at least 7,500. They are not of extreme length in any direction and
include no natural boundaries
5.) 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

2-12 LAUS Program Manual

Sample Design

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 of Primary Sampling Units
The most populous PSUs in a State are generally included in the CPS with
certainty. These are called self-representing (SR) PSUs. The remaining PSUs in a
State are called non-self-representing (NSR) PSUs. A probability sample of these
NSR PSUs is selected for the CPS from defined strata. There are 428 SR PSUs
and 326 NSR PSUs in the 1997 CPS sample.
The non-self-representing PSUs in a State are stratified and one PSU is sampled
per stratum. Stratification refers to the technique of splitting a larger population
(the State) into smaller subpopulations which are sampled separately. The number
of strata formed varies by State, and ranges from 1 in the District of Columbia to
42 in Texas. Information on PSU population, travel costs for data collection, and
other costs is used to determine the optimum number of strata. The NSR PSUs in
a stratum do not have to be contiguous. An algorithm is used to place “similar”
NSR PSUs in the same stratum, then any one PSU can appropriately represent the
entire stratum. A variety of demographic and economic variables are used to
stratify the PSUs, and these differ by State. The basic variables used in most States
are: male unemployed, female unemployed, families with a female head of
household, and proportion of households with 3 or more persons.
A typical stratum has 2-6 non-self-representing PSUs. The one NSR PSU chosen
from a stratum represents not just itself, but the entire stratum.
The probability of selecting a particular PSU in a non-self-representing stratum is
proportional to its 1990 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 chosen NSR PSU represents the stratum for the entire decade of the sample
design. A technique called “controlled selection” is actually used to maximize the
overlap of NSR PSUs between designs.

Selection of Households Using Census Data
Because the sample design is State-based, the sampling ratio differs by State and
depends on State population size as well as national and State reliability
requirements. The State sampling ratios range roughly from 1 in every 100

LAUS Program Manual 2-13

Sample Design

households to 1 in every 2,500 households. The sampling ratio used
within a sample PSU depends on the probability of selection of the PSU
and the sampling ratio for the State.
The 1990 within-PSU sample design uses census block level data from
the 1990 decennial census. A census block is a geographic area, patterned
after a “normal” city block, which encompasses approximately 40
housing units. Normally, census blocks are bounded by streets and other
prominent physical features such as rivers or railroad tracks. County,
Minor Civil Division, 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, census blocks are grouped into three
strata: “units”, “group quarters”, and “area”. The unit stratum 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 stratum contains housing units where
residents share common facilities or receive formal or authorized care or
custody. The area stratum contains blocks with addresses that are more
difficult to locate. Area blocks exist primarily in rural areas.
These strata 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 within-PSU sample reflects the
demographic and socioeconomic characteristics of the PSU, blocks within
the unit, group quarters, and area strata are sorted using geographic and
block-level data from the census. Examples of the census 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, and systematically sampling these strata, 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 withinPSU variance, compared to the variances associated with a simple random
sample of units within the PSU.
A USU is a cluster of four, mostly contiguous, housing unit addresses
which come into the sample and are interviewed at the same time. It is
more efficient, in terms of travel time and cost, to sample clusters of
housing units, as opposed to sampling individual units.

2-14 LAUS Program Manual

Sample Design

Units in the three strata described above all existed at the time of the 1990
decennial census. A sample of building permits, collected by the States in an
ongoing cooperative procedure, is included in the CPS to represent housing units
built after the decennial census. Adding these newly-built units keeps the sample
up-to-date and representative of the population. It also helps to keep the sample
size stable. Over the life of the sample, the addition of newly-built housing units
compensates for the loss of “old” units which may be abandoned, demolished, or
converted to nonresidential use.

CPS State Sample Sizes and Sampling Ratios
The CPS sample 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 100 to 1 in every 3,000 households in
each stratum of the State. The sampling ratio is occasionally modified slightly to
hold the size of the sample relatively constant given the overall growth of the
population. (This is called “sample maintenance reduction”.) 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 preliminary sample size estimate for households is
adjusted by a factor which increases the sample size to account for the normal
sample loss of eligible households that occurs due to household vacancies,
buildings demolished, etc.
The table below provides CPS sample sizes, assigned and eligible, for each State,
Los Angeles PMSA, balance of California, New York City, balance of New York
State, the District of Columbia, and the United States, as of January 1996.

LAUS Program Manual 2-15

Sample Design

CPS Sample Sizes, as of January 1996
Area

Assigned

Eligible

Area

Assigned

Eligible

United States

59,181

50,052

Missouri

778

668

Alabama

845

717

Montana

990

777

Alaska

887

662

Nebraska

823

710

Arizona

962

745

Nevada

722

616

Arkansas

874

732

New Hampshire

682

517

California

4,573

4,055

New Jersey

1,706

1,501

Los Angeles PMSA

1,828

1,626

New Mexico

894

695

Balance of California

2,745

2,429

New York

3,869

3,307

Colorado

824

713

New York City

1,691

1,472

Connecticut

618

550

Balance of New York

2,178

1,835

Delaware

678

545

North Carolina

1,594

1,297

District of Columbia

797

662

North Dakota

862

691

Florida

3,051

2,468

Ohio

2,144

1,924

Georgia

999

849

Oklahoma

1,027

822

Hawaii

670

560

Oregon

763

653

Idaho

940

740

Pennsylvania

2,574

2,193

Illinois

2,270

2,005

Rhode Island

678

566

Indiana

821

707

South Carolina

777

627

Iowa

813

710

South Dakota

915

760

Kansas

828

709

Tennessee

803

681

Kentucky

844

713

Texas

2,761

2,333

Louisiana

891

731

Utah

729

629

Maine

845

607

Vermont

769

556

Maryland

784

628

Virginia

938

823

Massachusetts

1,382

1,177

Washington

871

724

Michigan

1,995

1,765

West Virginia

960

797

Minnesota

864

732

Wisconsin

869

739

Mississippi

847

689

Wyoming

956

725

2-16 LAUS Program Manual

Sample Design

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 3,000, a within-PSU sampling ratio of 1 in 300
achieves the desired overall ratio of 1 in 3,000 for the stratum.

The Sample Rotation Design
The best estimates of month-to-month change would be obtained by surveying the
same households each month, or 100-percent sample overlap. However,
indefinitely surveying a single sample of households could lead to respondent
“fatigue” or “exhaustion” and the probability of increases in refusals and in
respondent errors.

% Identical HHs

Therefore, part of the sample is changed each month. Each monthly sample is
divided into eight representative subsamples, 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. 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 State-estimation modeling process. (See Chapter 6.)

Overlap of Identical Households
80 75
60
50
50
38
38
40
25
25
25
13
13
20
0 0 0 0 0
0
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Months between samples

LAUS Program Manual 2-17

Sample Design

The rotation plan used for the CPS sample also introduces nonsampling
error, referred to as month-in-sample bias. This bias generally refers to
the observed phenomenon of rotation groups differing, when theoretically
they should have approximately equal measurements. The 4-8-4 rotation
pattern adds an additional dimension to the month-in-sample bias because
of factors related to the large overlap of households from one month to the
next and one year to the next. Samples with large numbers of overlapping
units should have a high degree of consistency with regard to interview
responses. However, in repeated CPS interviews, the later interviews
yield consistently higher or lower estimates than earlier interviews.
Unemployment data exhibit the most pronounced month-in-sample bias,
with various subgroups, such as nonwhites and females, exhibiting more
bias than the general population. Historically, the following national
trends have been fairly regularly demonstrated:
• The unemployment rate drops from month-in-sample one to month-insample two and from month-in-sample five to month-in-sample six.
• There is an overall trend for the rate to decline from month-in-sample
one to month-in-sample eight.
• There is an increase, or surge, in the unemployment rate from monthin-sample three to month-in-sample four and from month-in-sample
seven to month-in-sample eight.
• There is general agreement that the month-in-sample four and eight
“surges” are attributable to the probing questions on discouraged
workers asked in those months. These questions apparently elicit
information that changes previous negative responses regarding the
“looking for work” questions, to positive responses. Thus, more
unemployed persons are identified. This probing was formerly done
in months-in-sample one and five, and significant changes in
reported responses result when the shift was made to months-insample four and eight.
It has also been observed that CPS sample responses vary from one month
to the next. There are a number of theories to explain this phenomenon.
One suggests that for a variety of reasons, sub-groups of potential
respondents are successfully interviewed at different rates. The degree of
differential response can change from one month-in-sample to another.
For example, suppose that employed persons living alone are harder to
find than other persons in month-in-sample one (since, often, no one is at
home when the interviewer calls). If arrangements are made with those
contacted in month-in-sample one to retain them in month-in-sample two,

2-18 LAUS Program Manual

Sample Design

then month-in-sample two could have a better representation of employed persons
living alone, relative to other persons, than month-in-sample one.
Because month-in-sample bias is believed to exist in the CPS, it is controlled for
in the compositing portion of the estimation process covered later in this chapter.
When the two components of the composite estimate are combined, a month-insample bias adjustment is added to adjust for the relative bias associated with
month-in-sample.

LAUS Program Manual 2-19

Data Collection

Data Collection
The housing units which belong to the selected USUs are called
“designated” households. The list of designated households is a
preliminary list of potential addresses to be sampled. Nationally, there are
approximately 59,000 designated households on this list. This list of
designated household units is then refined by adding households found by
reviewing building permits and 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 “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. There are
approximately 50,000 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 1,650 interviewers.
Personal visits are required in the first and fifth months that the household
is in the sample. If no one is at home when the interviewer visits, the
respondent may be contacted by telephone after the first month.
Approximately 60 percent of the households in any given month are
interviewed by telephone. About 12 percent of the households are
interviewed via the Computer Assisted Telephone Interview (CATI)
system from the CATI Collection Centers in Hagerstown, Maryland;
Jefferson, Indiana; and Tucson, Arizona.
On the first visit, the interviewer prepares a roster of the household
members and completes a questionnaire for each person 16 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

2-20 LAUS Program Manual

Data Collection

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.
Each month about 50,000 occupied units and approximately 100,000 individuals
are eligible for interview, and information is obtained for about 94,000 persons
aged 16 and older. On average, between 6 and 7 percent of these households are
contacted but interviews are not obtained because the occupants are not at home
(after repeated calls) or are unavailable for other reasons. These are called “Type
A” noninterviews.

Training and Quality Control
The sophisticated computer-based technology for the CPS requires specific
training and quality control procedures. CPS interviewers receive a combination
of home study, classroom, and field training, with monitoring and periodic review
and evaluation by CPS supervisory staff. Initial training includes fifteen hours of
home study, five days of classroom training sessions with mock interview practice
sessions, additional home study exercises during the first few months in the field,
supervised field data collection on selected days, and closely monitored official
data collection. Ongoing training includes two day-long classroom refresher
training sessions per year, monthly home study exercises, and one annual
observed field data collection (by supervisory staff). Staff at the CATI data
collection centers are subject to random monitoring of interviewing by
supervisory staff.
The data collection technology and the questionnaire provide an opportunity to
build functions to assist and improve data quality into the system itself. The
automated skip pattern not only allows more complex relationships between
questions, but also eliminates the chance that an interviewer might inadvertently
follow an incorrect skip path. 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 a “real-time” comments file stored in the computer,
the Census Bureau staff review and edit the information obtained for each person
in the sample, and, where possible, identify and correct omissions, unintelligible
entries, and other errors.

LAUS Program Manual 2-21

Data Collection

Quality control procedures also include monitoring “on line” CATI
interviews by Census Bureau supervisory staff; a system of reinterviews
where about 5 percent of the sample is reinterviewed and responses
compared with initial interview responses; and monthly feedback to the
field staff on any errors, omissions, or inconsistencies detected by the
computer edits.

2-22 LAUS Program Manual

Estimation Procedures

Estimation Procedures
There are six main steps to the estimation process in the CPS; editing of raw data
and imputation, basic weighting, noninterview adjustment, ratio adjustment,
compositing estimates, and seasonal adjustment. This process takes the raw data
from the CPS interviews, edits it, weights it to represent the population as a whole,
adjusts the data 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 is completed by Census staff at a
central location in Suitland, MD, and 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 national Census Bureau, they are once again 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 from the most recently processed record for a person in the
same age/sex/race/geography cell.
Basic Weighting

The basic weighting procedure begins the process of inflating the sample data to
produce an estimate for the entire population. In the basic weighting procedure,
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 unbiased estimate of the
number of persons in the population possessing that characteristic.

LAUS Program Manual 2-23

Estimation Procedures

When a selected cluster of housing units is found to have many more units
than expected, field subsampling is carried out. Appropriate special
weights, reflecting the subsampling of the cluster, are then applied to the
sample data.
Noninterview 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 noninterview
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 Statistical Area (MSA) status and
size. Within each cluster, there is a further breakdown by residence. Each
MSA cluster is split by “central city” and “balance of the MSA”. Each
non-MSA cluster is split by “urban” and “rural” residence categories.
The proportion of sample households not interviewed averages between 6
and 7 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.
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.
It is applied only to PSUs that are non-self-representing in States
that have a substantial number of black households
(approximately 20 States). 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

2-24 LAUS Program Manual

Estimation Procedures

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.

The second-stage ratio adjustment (also known as “raking”) is carried out in three
basic steps. In the first step, the sample population and the labor force estimates
are adjusted within each State and the District of Columbia using an independent
control for the population 16 years and over. This forces the CPS State estimate to
equal the independent State estimate. The second adjustment involves an
adjustment by Hispanic origin to a national estimate for 14 Hispanic and 5 nonHispanic age-sex cells. In the third step, a national adjustment is made by the race
categories of white, black, and other races to independent estimates by age and
sex. The white and black categories contain 66 and 42 age-sex groups,
respectively; the other races category has 10 age-sex cells. The entire secondstage ratio adjustment procedure is iterated six times, each iteration beginning
with the weights developed during the previous iteration. This insures that the
adjusted sample population estimates for both the States and the national age/sex/
race/Hispanic origin categories will be virtually equal to the independent
population controls for these categories.
The monthly independent State controls for the civilian noninstitutional
population 16 years and over in the raking process are based on an arithmetic
extrapolation of the trend in population growth (or decline) using the two most
recent July 1 estimates, with all State estimates prorated to a current estimate of
the U.S. population. The projections are derived by updating demographic census
data with information from a variety of other data sources that account for births,
deaths, and net migration. The National Center for Health Statistics (NCHS)
provides the Census Bureau with data on births by age, sex, race, and Hispanic
origin, although data for the latest month must be projected. Deaths by age, sex,

LAUS Program Manual 2-25

Estimation Procedures

and race are also provided by NCHS, although the latest 6 months must be
projected from a life table based on NCHS and Social Security
Administration data. (The entire series of deaths for the Hispanic-origin
population is projected.) Data on legal international immigration are
obtained from the Immigration and Naturalization Service, the Puerto
Rican Planning Board, and the Office of Refugee Settlement. Estimates
of net undocumented immigration and permanent emigration of legal
United States residents are modeled using data from surveys and other
censuses. The net movement of U.S. citizens from overseas to the United
States is estimated based on data provided by the Department of Defense
and the Office of Personnel Management (for military and civilian
Federal Government personnel and their dependents). Other net
migration is assumed to be zero. Most of the data are characterized as
administrative, although some data for recent months must be projected.
Thus, while the data are not subject to sampling error, they may contain
nonsampling error and bias. Estimates of net census undercount,
determined from the Post Enumeration Survey, are added to the
population projections.
Composite Estimation

The last step in the preparation of most CPS estimates makes use of a
composite estimation procedure. Statistical theory states that the estimate
of a quantity can be improved if two (or more) estimates of that quantity,
obtained by different methods, are combined. This technique is called
compositing.
The composite estimate consists of a weighted average of two estimates:
1.) The second-stage ratio estimate based on the entire sample from
the previous month. (This estimate receives a weight of 60
percent.)

and
2.) An estimate for the current month obtained from the prior
month’s final (composited) estimate updated by adding an
estimate of over-the-month change based on the sample
common to both months (about 75 percent of households). (This
estimate receives a weight of 40 percent.)

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.

2-26 LAUS Program Manual

Estimation Procedures

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 sample overlap of the survey. For some items, the reduction is substantial.
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 interval of time.
Seasonal Adjustment

Seasonal events, such as weather changes, harvests, major holidays, and school
openings and closings cause fluctuations in employment and unemployment
levels. Seasonality, which may account for as much as 95 percent of month-tomonth unemployment change, obscures nonseasonal trends and cyclical
movements. Since seasonal fluctuations follow fairly regular annual patterns,
their influence can be eliminated from data series through seasonal adjustment.
The X-11 ARIMA procedure, a process employed by the CPS, is based on the
standard ratio-to-moving average method from time series analysis.
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. (See
Chapter 6 for a more complete discussion of seasonal adjustment.

LAUS Program Manual 2-27

Estimation Procedures

2-28 LAUS Program Manual

Unemployment

Insurance System

3

Inputs to LAUS Estimation:
The Unemployment Insurance
System

he 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. In addition to the State UI programs which cover the bulk of nonfarm
workers, separate Federal UI programs exist for railroad workers through the
Railroad Retirement Board (RRB), for Federal employees through
Unemployment Compensation for Federal Employees (UCFE), and for exservicemen through the Unemployment Compensation for Ex-Servicemen
(UCX).

T

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.
Statistics from the UI systems are the only current measure of unemployment at
the substate level available at the county (and in some States, city) level. They are
the major input to the Unemployment Rate models used to estimate
unemployment for the 50 States, the District of Columbia, Los Angeles-Long
Beach, the balance of California, and New York State minus New York City. (The
UI claims series for New York City did not provide a reliable predictor of the
unemployment level and are, therefore, not included in the regression model for

LAUS Program Manual 3-1

New York City.) Claims data from the UI systems are inputs to the
Handbook method for estimating LMA unemployment, and their use in
the claims-based unemployment disaggregation yields more accurate subLMA estimates than are obtained from decennial census-based
approaches.
While these statistics are biased for estimating total unemployment
between States in so far as they reflect the particular State’s UI law and
administrative practices, the statistics have the advantage of being current
and, with proper coding and tabulation, are consistent among areas within
States.

3-2 LAUS Program Manual

State Role

State Role
States are responsible for the following UI activities.
1.) Claims-taking. The State UI offices accept initial claims information from
individuals who are filing for benefits. This information may be obtained
in person from the claimant in an Employment Service office,
electronically from the claimant entering information in an Employment
Service kiosk established for claims-taking, or over the telephone.
2.) Monetary eligibility determination. In accordance with the State’s laws,
the State determines if the individual is covered by the UI system and, if
so, how much in benefits the individual is eligible to receive.
3.) Nonmonetary determination. The State disqualifies ineligible individuals
from receiving benefits based on nonmonetary issues. These include the
circumstances surrounding the loss of employment, ability to work,
availability for work, and activity in seeking work.
4.) Benefits delivery. The State provides benefit payments to the
unemployed individuals who successfully certify to a week of
compensated unemployment. Benefits may be delivered under any of the
following arrangements:

• Intrastate Benefit Arrangements. The State provides benefits to
individuals who reside 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 (contiguous) State.
• 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 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
electronic means. Determinations on eligibility, disqualifications, and
the amount and duration of benefits are made by the liable State.
• Wage-Combining Arrangements. This arrangement permits multi-State
workers to combine their wages and employment in more than one
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

LAUS Program Manual 3-3

State Role

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.

3-4 LAUS Program Manual

UI Coverage

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
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.

LAUS Program Manual 3-5

UI Coverage

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 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, but are not included in ES-202 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 ES-202 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.

3-6 LAUS Program Manual

Differences: UI Data versus the CPS

Differences: UI Data versus the CPS
CPS data are used directly to produce official labor force estimates for the nation.
According to CPS concepts, a person who did any work at all for pay (or at least
15 hours unpaid in a family business) during the survey reference week (generally
the week including the 12th of the month) is employed. 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 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.
Most States require a waiting period before individuals can receive benefits. A
waiting period is a noncompensable period of unemployment in which the
individual must otherwise have been eligible for benefits. Typically, the waiting
period is one week.

LAUS Program Manual 3-7

BLS Standards for UI Data

BLS Standards for UI Data
UI Data Standardization: The UI Database Survey
Beginning in 1975 an
effort was undertaken by
BLS to survey the State UI
database systems. There
were two primary reasons
for the survey. The first
was to determine the
nature of the data obtained
for LAUS purposes, and
the second was to contract
with States to modify their
systems where necessary
to achieve more uniform
data series.

The UI
Database

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 UI claimant statistics used in unemployment
estimation, and are areas where improvement efforts have been
concentrated.
To obtain an unduplicated count of persons, Social Security numbers are
used. Also, the use of the week of certification precludes multiple
counting because an individual can certify only once to a week of
unemployment.
Basing the UI statistics on claimant residency rather than more programrelated 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.

3-8 LAUS Program Manual

BLS Standards for 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 make the data consistent with the
CPS, where one hour of pay qualifies an individual as employed.

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, through
disaggregation, of county and subcounty estimates.
State and County of Residence

The residency requirement for claims data calls for the coding and tabulating of
claimants by State and county of residence. The geographic requirement applies
to counties 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). If State claims documents (either intrastate or
interstate) are preprinted with the State code, border State codes must be entered
for commuter claimants to insure correct residency information.
Commuter Claimants

Commuter claimants are individuals who worked and would continue to seek
work in one State while living in another State. These claimants are treated as if
they reside in the State of employment and file intrastate initial claims in the State
in which they had worked.
Interstate Claimants

Interstate claimants are individuals who file claims for compensation either
through the facilities of an agent State (usually their State of residence) or directly
to the liable State (the State in which the last employer is located) via the
telephone or electronically.

Continued Claimants and Final Payment Recipients
Two insured unemployed counts, continued claimants and final payment
recipients, are used in the development of Handbook LMA unemployment
estimates and for the model-based unemployment estimation procedure.
Continued claimants are persons certifying to a compensated or noncompensated
week of unemployment under the State UI and UCFE programs. Because

LAUS Program Manual 3-9

BLS Standards for UI Data

measurement is limited to the labor force status of the civilian population,
the UCX program is excluded. Continued claimant groups include
intrastate claimants, commuter claimants (based on State of residence),
and interstate claimants (based on State of residence).
Persons receiving final payments are continued claimants certifying to a
compensated week of unemployment which represents the last regular
benefit payment in the benefit year because no further benefits are
available until the beginning of a new benefit year.
The BLS standard of quality for these two insured unemployment counts
is as follows:
• the counts reflect the State and county of residence of the unemployed;
• the counts are unduplicated and based on the week of unemployment
for which the claimant certified;
• 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);
• for persons receiving final payments, the counts are weekly, based on
the week for which the claimant is certifying; and
• the counts exclude persons with any earnings due to employment,
regardless of their entitlement to full weekly UI benefits.

Standards for Initial Claims
Layoff!

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 no longer required in
Handbook estimation, but it is
useful to define the standards for
initial claims.)

An initial claim is a notice filed 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

3-10 LAUS Program Manual

BLS Standards for UI Data

after an intervening period of employment. The initial claims count which can be
used in LAUS estimation includes both new and additional initial claims filed for
State UI in the week including the 19th of the month.
Reference Period

Unlike continued claims which 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

The following types of initial claims are excluded from consideration:
• Invalid new initial claim. An initial claim where the individual is found to be
monetarily ineligible for UI.
• Transitional initial claim. An initial claim filed by a claimant during a spell of
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.
• Reopened claims. A claimant may cease certifying to unemployment and
withdraw 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.
Residency

Since the initial claims count may be an input into the Handbook estimate at the
area level, the count must be tabulated by the State and county of residence of the
claimant.

Standards for Nonmonetary Disqualification
Nonmonetary disqualification may be used in atypical or exception situations in
the Handbook procedure to estimate the unemployed who are disqualified from
receiving benefits but still meet the definition of unemployed in the CPS.
Unemployed Disqualified

All States disqualify claimants who voluntarily quit without good cause, who are
discharged for misconduct, who refuse an offer of suitable work without good
cause, as well as those whose unemployment results from a labor dispute. In

LAUS Program Manual 3-11

BLS Standards for UI Data

addition, individuals must also participate in reemployment services, such
as job search assistance, if he/she is determined through a profiling
system as likely to exhaust regular benefits. But, the definition of key
terms like “good cause”, “misconduct”, and “suitable work” may differ
from one jurisdiction to another.
In the UI system, a person not complying with requirements to be able,
available, and actively seeking work will be disqualified from receiving
benefits. In the CPS, if a person is not able to work or is not available for
work, that person is out of the labor force, as is a person who has not
actively sought a job in the previous four weeks. Therefore, in general,
the outcome of unavailability for work is the same under UI and the CPS,
that is, out of the labor force. However, in the UI system, the requirement
is more stringent in terms of weekly activity.
In the UI system, a claimant will be considered unemployed and will
receive benefits only if the claimant lost the job through no fault of his/her
own. Therefore, people who voluntarily quit or are discharged for
misconduct will not receive UI benefits. However, these individuals are
counted as unemployed in the CPS.
Issues

In order to estimate the unemployed who are disqualified from receiving
UI benefits only separation issue denials should be used. This count
encompasses disqualifications for voluntary separation, termination for
misconduct, and special statutory requirements including issues based on
leaving a job, misconduct, and other issues.
Denials involving issues whereby disqualified claimants would be
classified as out of the labor force are not used. This applies to denials for
able, available, and actively seeking work requirements. Denials for
refusal of suitable work are viewed as voluntary withdrawals from the
labor market on the part of the disqualified claimants. Denials for
disqualifying income are not used because of problems in differentiating
between short-term postponement penalties and penalties of reductions in
benefits which allow the individual to receive a reduced check. In the
latter case, the claimant would appear in most cases as a continued
claimant but with earnings. For the former, problems exist with the
identification of the reference week in short-term disqualifications.
Reference Period

The basis for the week designation may be the week in which the
determination was made, or the effective week of the penalty
(disqualification), as defined by the State UI law. In most cases, the week

3-12 LAUS Program Manual

BLS Standards for UI Data

of the imposition of the penalty is used, due to the extensive revisions that would
be required by using other determinations, such as the effective week of the
penalty.
Types of Penalties

The penalties imposed on disqualified claimants vary considerably among the
States. They may include one or a combination of the following: a postponement
of benefits for some prescribed period, ordinarily in addition to the waiting period
required of all claimants; a cancellation of benefit rights; or, a reduction of
benefits otherwise payable. Disqualification means that benefits are denied for a
definite period or for the duration of the period of unemployment.
The disqualification period may be for the week of the disqualifying act and a
specified number of consecutive calendar weeks following, or for the duration of
unemployment, or longer by requiring a specified amount of work or wages to requalify. The theory of a specified period of disqualification is that, after a time,
the reason for a worker’s continued unemployment is related more to the general
conditions of the labor market than his/her disqualifying act.
Residency

Since the separation issue disqualification count may be an atypical input to the
Handbook estimate at the area level, the count should pertain to the State and
county of residence of the claimant.

LAUS Program Manual 3-13

BLS Standards for UI Data

3-14 LAUS Program Manual

Establishment

Data Sources

4

Inputs to LAUS Estimation:
Establishment Data Sources

here are two establishment-based data sources for employment estimates.
These are the Current Employment Statistics (CES) program and the
Covered Employment and Wages Program, commonly referred to as the
ES-202 program. The next two sections provide an overview of these two
programs.

T

The Current Employment Statistics Program
The Current Employment Statistics (CES) survey
is a Federal-State cooperative survey of
approximately 400,000 business establishments
nationwide. The program produces monthly
employment estimates for the nation, each State,
and 272 of the 334 Metropolitan Statistical Areas
(MSAs) 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, or store, which produces goods or provides services. It is
generally at a single physical location and engaged in one, or predominantly one,

LAUS Program Manual 4-1

The Current Employment Statistics Program

type of economic activity. Where a single location encompasses two or
more distinct activities, these are treated as separate establishments,
provided that separate payroll records are available.
Employment. Employment is the total number of persons employed fullor 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. For Federal government establishments,
employment statistics relate to civilian employees only and are reported
for the number of persons who occupied positions on the last day of the
calendar month. 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 combines annual benchmarks
from the Covered Employment and Wages program with monthly data
from a sample survey to produce estimates of employment, hours, and
earnings. All firms with 250 employees or more are asked to participate
in the survey. A sample of smaller firms, with probability of selection
proportionate to size, is also selected. The sample frame is the master list
of establishments reporting to the Unemployment Insurance system and
maintained as the Universe Maintenance System by BLS.
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. This amounts to distributing the total

4-2 LAUS Program Manual

The Current Employment Statistics Program

number of establishments needed in the sample among the cells according to the
ratio of the employment in each cell to the total employment in the industry.
Data are collected from the establishments surveyed on the report form BLS 790
or electronic equivalent. (The CES survey is often referred to as the 790
program.) CES estimates are made at a basic estimating cell level and then
aggregated to industry total levels by simple addition. The CES employment
estimates at the cell level are derived by the following steps:
Step 1.

A total annual benchmark is taken from ES-202 data.

Step 2.

Employment data are gathered from a CES sample for the current
month.

Step 3.

A ratio of all employees from the current month to those in the previous
month for each cell is computed from the sample of establishments
reporting for both months. This ratio is called the “link relative”.

Step 4.

The final all-employee estimate from the previous month is multiplied
by the link relative for the current month. This process begins with the
benchmark month and moves forward to the next benchmark.

In some States, and for the national estimates, a bias adjustment factor to account
for new business births during the month is applied to the estimate for the current
month creating a final all-employee estimate for the current month.
This method, the “link-relative technique”, produces month-to-month changes for
a matched sample of industry establishments. Aggregate monthly estimates are
produced by industry and geographic area.
Benchmarks

The establishment survey constructs annual benchmarks in order to realign the
sample-based employed totals for March of each year with the UI-based universe
counts for March. These population counts provide an annual point-in-time
census for employment.
Universe counts are derived from the administrative file of employees covered by
UI. Approximately 99 percent of in-scope private employment is covered by UI.
A benchmark for the remaining 1 percent is constructed from alternate sources,
primarily records from the Interstate Commerce Commission and the Social
Security Administration. The full benchmark developed for March replaces the
March sample-based estimate for each basic cell. The monthly sample-based
estimates for the year preceding and the year following the benchmark are also
then subject to revision.

LAUS Program Manual 4-3

The Current Employment Statistics Program

Monthly estimates for the year preceding the March benchmark are
readjusted using a “wedge back” procedure. The difference between the
final benchmark level and the previously published March sample
estimate is calculated and spread back across the previous 11 months.
The wedge is linear; eleven-twelfths of the March difference is added to
the February estimates, ten-twelfths to the January estimates, and so on,
back to the previous April estimates which receive one-twelfth of the
March difference. This assumes that the total estimation error since the
last benchmark accumulated at a steady rate throughout the current
benchmark year.
Estimates for the 11 months following the March benchmark are also
recalculated each year. These post-benchmark estimates reflect the
application of sample-based monthly changes to new benchmark levels
for March, and the recomputation of bias adjustment factors for each
month. Bias factors are updated to take into account the most recent
experience of the estimates generated by the monthly sample versus the
full universe counts derived from the UI.
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 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.

4-4 LAUS Program Manual

The Covered Employment and Wages Program

The Covered Employment and Wages Program
Background
The Covered Employment and Wages program, commonly called the ES-202
program, is a cooperative endeavor of BLS and the employment security agencies
of the 50 States, the District of Columbia, Puerto Rico, and the Virgin Islands.
Using quarterly data submitted 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 ES-202 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 ES-202 program is quarterly and
provides a virtual census of nonagricultural employees and their wages. In
addition, about 44 percent of all workers in agricultural industries are covered.

Sources of Data
There are five sources of data for the ES-202 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 ES-202 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.

LAUS Program Manual 4-5

The Covered Employment and Wages Program

Federal Government Reports

These reports are filed quarterly by most federal government agencies to
report employment and wages data to the State ES-202 unit, in
accordance with the Unemployment Compensation for Federal
Employees (UCFE) program. Data for non-defense federal agencies are
provided to the State ES-202 unit; information for civilian employees of
the Department of Defense are reported directly to BLS-Washington.
Annual Refiling Survey Forms

UI-liable employers are surveyed by the State ES-202 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 five ownership
categories into which establishments are classified are private industry,
Federal government, State government, local government, and foreign or
international 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 Standard Industrial
Classification Manual 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.

4-6 LAUS Program Manual

The Covered Employment and Wages Program

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
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. Workers on intrastate and
scenic railroads may be covered.

LAUS Program Manual 4-7

The Covered Employment and Wages Program

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, but are not included in ES-202 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 ES-202 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.

4-8 LAUS Program Manual

The Covered Employment and Wages Program

Uses
The ES-202 data serve as the basic source of benchmark information for
employment by industry and by size of establishment in the Current Employment
Statistics program.
The Unemployment Insurance Name and Address File, developed in conjunction
with the ES-202 report, also serves as a national sampling frame for establishment
surveys by the Producer Price Index, Occupational Safety and Health Statistics,
Employment Cost Index, and other compensation programs.

LAUS Program Manual 4-9

Differences: Establishment Data Sources versus the CPS

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, others
cannot. It is useful to be aware of the CPS/CES/ES-202 differences for
estimation and analysis purposes. Some of the important differences are
discussed below.
Place of Work vs. Place of Residence. CES and ES-202 data are produced
according to the location of the establishment; CPS data provide
residency-based employment estimates.
Jobs versus Employed People: Workers holding more than one job may be
included more than once in the CES and ES-202 employment counts
since they may appear on more than one payroll record or contribution
report. Persons counted by 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 5th. The reference period for CES and ES-202 is the payroll
period including the 12th of each month, which could be weekly, biweekly, semi-monthly, etc.
Employment Coverage Differences. The CPS definition of employment
comprises 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 familyoperated enterprises. Employment in both agricultural and nonagricultural
industries is included. The CES and ES-202 do not include self-employed
and unpaid family workers. They include some, but not all, domestics in
private households and agricultural workers.
CES and ES-202 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

4-10 LAUS Program Manual

Differences: Establishment Data Sources versus the CPS

interview. Workers who are on strike for the entire pay period of the establishment
are not included in the CES and ES-202 estimates, but are considered employed in
the CPS.

LAUS Program Manual 4-11

Uses of CES Data in the LAUS Program

Uses of CES Data in the LAUS Program
The nonagricultural wage-and-salary estimates from the CES survey are
used as basic employment inputs for several LAUS estimating
procedures. CES estimates are used as variables for the State employment
models. CES data are used in adjusting place of work employment to
place of residence and as current inputs to labor market area employment,
where available.

Use of CES in the State Employment Models
The statewide CES employment is a
variable in the State employment/
population model. In the model, the CES
is used as the major data source for the
employed portion of the labor force. The
CES data are adjusted to include
individuals involved in labormanagement disputes. Those differences between the CPS and CES
discussed in the previous section do not require separate adjustments in
the model because coefficients are computed separately for each State
from their own data and the data relationships. The coefficients represent
data relationships and are used to account for differences between the two
data sources. The CES variable is always included in the employmentpopulation model because nonagricultural wage and salary workers
represent such a large proportion of the employed. See Chapter 6,
Statewide Estimation, for more information on modeling techniques.

Use of CES for Estimating Current LMA Employment
Labor Market Areas (LMAs) which are officially participating in the CES
program use the nonagricultural wage and salary CES estimates in
developing monthly LAUS total employment estimates. If a LMA is not
covered by the CES program, but does have a sample-based employment
series developed under State auspices, these estimates are used in the
Handbook methodology. For small LMAs without sample-based
employment estimates, nonsample (synthetic) estimation methods, using
covered employment and wages estimates, are used to yield place-ofwork nonfarm employment. See Chapter 7, for details on producing
estimates for areas outside the CES program.

4-12 LAUS Program Manual

Uses of CES Data in the LAUS Program

Use of CES in the Residency-Adjustment Ratio
Current monthly area nonfarm employment estimates, which are establishmentbased, are converted to residency based by the application of the Census/790 ratio.
This ratio, also known as the Total Nonagricultural Wage and Salary Residency
Adjustment Ratio, is computed for each area as:
The area’s decennial census estimates of total employment divided by the area’s
CES (790) employment estimate for March/April 1990.

When multiplied by the current month’s sum of CES total nonagricultural wage
and salary employment and the number of labor disputants, this ratio yields the
residency-based LAUS total nonagricultural wage and salary employment.
See Chapter 7 for a more detailed description of this adjustment.

LAUS Program Manual 4-13

Uses of CES Data in the LAUS Program

4-14 LAUS Prorgam Manual

Decennial

Census

5

Inputs to LAUS Estimation:
Census Data

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. The decennial census also has increasingly
become a source of data for other uses and provides
socioeconomic and demographic data in addition to
population estimates.

T

The LAUS program methodology uses decennial census data for adjusting
establishment-based employment estimates to residency-based employment
estimates, for estimating certain employment and unemployment components in
the Handbook methodology, and disaggregating or apportioning labor market area
estimates to smaller areas.

The Decennial Census: Enumerated and Sample-Based
Data
General Description of the Census Questionnaire and
Resulting Tabulations
The census questionnaire comes 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 receive a short form
where questions regarding household relationships, sex, race, age, marital status,

LAUS Program Manual 5-1

The Decennial Census: Enumerated and Sample-Based Data

Hispanic origin and housing are asked. Approximately one-sixth of the
population receive 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.
Tabulations based on the 100-percent enumerated questions are prepared
down to the block level. Tabulations for the sample questions are also
prepared down to the block level, but because they are based on a sample,
the data are reliable only for larger areas. Areas for which statistics are
derived include Census regions and divisions, Metropolitan Statistical
Areas, Urbanized Areas, Urban/Rural areas, Census county divisions,
Census Designated Places, Census tracts, Block Numbering Areas, Block
groups, Blocks, Alaska Native village statistical areas, Tribal designated
statistical areas, and Tribal jurisdictional statistical areas.

Sampling for Designation of Long Form Recipients
The basic sampling unit for the long form census is the housing unit,
including all occupants. For persons living in group quarters, the sampling
unit is the person. Two sampling rates are employed. In counties,
incorporated places, and minor civil divisions estimated to have fewer
than 2,500 persons, one-half of all housing units and persons in group
quarters are included in the sample. In all other places, one-sixth of the
housing units and persons in group quarters are sampled. The purpose of
this sample design is to provide reliable estimates for small places. When
both sampling rates are taken into account, approximately 19 percent of
the housing units in the nation are included in the sample.

Estimation of Census Sample Data
The estimation procedure used for the sample survey involves an iterative
ratio estimation technique, called proportional fitting, similar in concept
to the estimation procedures used for the Current Population Survey. In
each tabulation area, a characteristic total is estimated by summing the
weights assigned to the persons or housing units in the area. Initial
weights for both households and persons are assigned as the approximate
inverse of the probability of selection for the Census sample. Weighting
areas are then created with a minimum sample of 400 persons.

5-2 LAUS Program Manual

The Decennial Census: Enumerated and Sample-Based Data

Within a weighting area, the ratio estimation procedure is conducted in four stages
for both persons and occupied housing units. The first stage identifies 17
household-type groups which include classification by the number of persons in a
house and type of housing unit (e.g., persons in a housing unit with or without
children, group quarters, etc.). The second stage determines the sampling rate of
the weighting area. The third stage classifies persons as household/nonhouseholder and housing units as single- or multiple-units in a structure. The
fourth stage applies 180 aggregate age/sex/race/Hispanic origin categories.
Groups within these four categories are combined, if needed, to increase
reliability. In the final step, the initial weights undergo four stages of ratio
adjustment by which each group within each stage is multiplied in two iterations
by the ratio of the complete census count to the sum of the initial and subsequent
stage weights for each sample person.
Sample data are considered less reliable than enumeration, or 100-percent
questionnaire, data. However, estimated standard errors can be used to construct
confidence intervals around the sample estimates. These reliability estimates do
not account for nonsampling errors that are inevitable in a survey as extensive as
the decennial census, and which occur in enumerated as well as sample-based
data.

Nonsampling Errors
Nonsampling error can occur in the enumerated and the sample-based data, and
can introduce bias into the data as well as increase the total error associated with
the estimates. The Census Bureau tries to control for such error during collection
and processing procedures. Types of nonsampling error include undercoverage,
respondent and enumerator error, processing error, and nonresponse.
Every census results in an undercount, i.e., some people are missed. These
undercounts can occur by age, sex, and race categories. The Census Bureau
compares its data to other aggregate data sources to analyze the demographic
count differences. It also conducts a post-enumeration survey by taking a sample
of areas within the US and doing a very accurate count of the persons in those
areas. This allows the Census Bureau to estimate the extent of undercount. For the
1990 census, the total undercount of the population was less than 2 percent.

LAUS Program Manual 5-3

Differences: Census versus CPS/LAUS Estimates

Differences: Census versus CPS/LAUS Estimates
Historically, there have always been differences between the Census and
the CPS programs in their respective estimates of unemployment. Prior to
1990, however, the census-based unemployment estimates generally
tended to be very close to the CPS estimates. For the 1990 census, this
relationship changed, with the 1990 census unemployment estimates
considerably higher than those from the CPS.
There are several important differences between methodology used in
CPS/LAUS and the methodology used in the decennial census. These
produce differing unemployment estimates. It is important to know how
to interpret these differences and explain why the CPS/LAUS estimates
are regarded as the more reliable and accurate estimates.
1.) The census is a self-enumeration survey, while the CPS survey is
conducted in an interviewer-controlled environment. This
provides the CPS with more accurate and detailed response
information because interviewers are present to clarify
questions.
2.) The CPS questionnaire asks seven specific employment-related
questions to arrive at the labor force classification. The census
questionnaire asks only four. Misclassification can occur as a
result of fewer employment-specific questions.
3.) The CPS has rigorous quality control procedures. Interviewers
are trained extensively, proficiency checks are conducted
regularly, and a portion of each month’s households are
reinterviewed as a quality control measure.
4.) The CPS has a definite reference period, i.e., generally the week
including the 12th of the month. The census reference period is
officially April 1, but the questionnaire instructs the respondent
to provide information as of the week before the questionnaire is
completed. In 1990, 75 percent of the census questionnaires
were completed by mid-May, however, some were not completed
until August. (This could have biased the estimates toward
higher unemployment because in August of 1990, the economy
had entered a recession.)
5.) There is a known first month-in-sample reporting bias whereby
unemployment rates tend to be higher the first time a household
reports information. In the CPS, households are interviewed for 4
months, not interviewed for 8 months, and then interviewed
again for 4 months, so that 25 percent of the sample could be
reflecting this bias. The census is a one-time survey.
Consequently, the entire census could be affected by first-time
reporting bias.

5-4 LAUS Program Manual

Differences: Census versus CPS/LAUS Estimates

Because of these reasons, both BLS and the Census Bureau have agreed that the
superior estimator of the labor force is the CPS.
The LAUS program uses decennial census data only where no other source of data
is available. The direct use of decennial census data is generally avoided because
of the superiority of the CPS and because relationships in the data are unlikely to
remain fixed over an entire decade.

LAUS Program Manual 5-5

Uses of Decennial Census Data in LAUS

Uses of Decennial Census Data in LAUS
Uses of Labor Force Estimates
Census employment estimates are used in the employment/population
index share disaggregation method, which is used in conjunction with the
claims-based unemployment disaggregation method for counties and
cities. Because more current estimates of employed residents are not
available, the decennial Census estimate of this group is moved over time
by changes in annually prepared population estimates. In other words, the
Census E/P ratio is maintained over the decade.
Decennial labor force estimates of total employed and unemployed for
sub-county areas are the basis of the Census-share disaggregation method.
The use of Census data for disaggregating labor force estimates is
required when UI claims data by county or city of residence are not
available. The method uses ratios of employment and unemployment in
subareas to the respective larger area total. In this method, the relative
distribution of employed and unemployed is fixed for the decade.
In order to develop place-of-residence employment estimates for LMA’s,
census nonfarm employment levels, in combination with CES payroll
employment levels, are used as residency adjustment factors for monthly
establishment-based employment estimates. (Net commutation patterns
are fixed.)
The census employment levels of agricultural employment and all-other
employment (self-employed, unpaid family workers, and domestics in
private households) are the benchmark levels for current LMA estimates
of these components.
The census estimates of all-other employment are also used in the
stratification of States into three groups for the purpose of developing
monthly change factors. The monthly change factors, referred to as Step-3
Ratios, are then used to estimate monthly all-other employment levels.
Census journey to work commutation data, which identify place of
residence and place of work estimates, are used in the designation of
LMA’s, including metropolitan areas and small labor market areas.

5-6 LAUS Program Manual

Uses of Decennial Census Data in LAUS

Uses of Population Data
Census total population data have been
used in the population-share
disaggregation method for determining
sub-county estimates. This method is
used only when subarea UI claims data or
census labor force data are not available.
The disaggregation is based on the ratio
of total population in a subarea to total
population in the larger area, applied to
current employment and unemployment estimates for the larger area. This method
is applied only after the claims-based/population-based disaggregation and
census-share methods have been used to establish estimates for the larger areas.
Approval from the appropriate regional office must be obtained before employing
this disaggregation method.
Decennial census population estimates for States, and the subsequent intercensal
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.
In the Handbook methodology, decennial Census State population counts by age
are used to establish age categories used in the calculation of Youth Population
Ratios (YPR) for the estimation of unemployed entrants and reentrants. In
subsequent years, survival rates are used below the State level to update the
YPR’s.
Age-specific population counts are also used in the claims-based unemployment
disaggregation of LMA entrants and reentrants. The age groups used in this
method, 16-19 and 20 and older, are consistent with those used in calculating the
YPR.

LAUS Program Manual 5-7

Uses of Decennial Census Data in LAUS

Employment/Unemployment Data
Data
Total Employment:

Use
Disaggregation of employment estimates

county and sub-county levels
Total Unemployment

Disaggregation of unemployment estimates

Employment:
for Selected SICs

Determination of appropriate weighting for combined residency-adjustment factor

Agricultural Employment

Agricultural employment benchmark

All-other Employment

Stratification of areas based on 1980/90 relative
change and domestics and self-employed/unpaid
family benchmark

Commutation Data

MSA and small LMA redefinition

Population Data
16+ Civilian, Non-institutional
Population for States

CPS population controls

Total Population by Age
Group for LMAs

Youth Population Ratios for estimating unemployed entrants

Total Population to SubCounty Levels

Disaggregation of employment from LMAs to
counties and counties to places

Total Population 16-19, 20+ to
Sub-county Levels (cities over
25,000)

Claims disaggregation of LMA unemployment estimates

5-8 LAUS Program Manual

Uses of Decennial Census Data in LAUS

Post-Censal Population Estimation
Post-censal population estimates are used in the State CPS estimation
methodology and the LAUS employment/population index share disaggregation
of the estimates for counties and, in some States, cites over 25,000. 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 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.

Statewide Estimation
National population estimates from the Bureau of the Census, which account
directly for births, deaths, and legal immigration, are done on a relatively
straightforward basis. The population of the States must be estimated using less
direct methods, because interstate migration, a large component of the change in
State populations, cannot be accounted for directly in the way that births, death,
and legal immigration can. There are two methods that are used, and the official
State estimates are an average of the two. The sum of the States is made to equal
the national total in a process known as proportional “raking”.
The first method for Statewide estimation uses available administrative records.
This method measure interstate migration for persons under age 65 by using
address changes on tax returns and the number of exemptions claimed. For the
population of age 65 and over, migration is measured using the change in the
number of Medicare enrollees.
The second method uses a composite of a number of different factors. They are:
• migration of the school-age population based on school enrollments;
• a regression on changes in various population indicators measurable from
various data sources, for the population of adults under 65 years old; and
• Medicare enrollments, for the change in the population 65 and older.
The age pattern of migration used in producing estimates of State populations by
age and sex (which affects the total 16 and over) depends primarily on the age
distribution of the base population, changes in school enrollments, and the
relationship between school-age migration and adult migration as of the last
census.

LAUS Program Manual 5-9

Uses of Decennial Census Data in LAUS

Substate Estimation
Substate population estimates are issued by the Bureau of the Census in
two series. In the P-25 series, Census publishes provisional estimates for
States, counties, MSA’s, and incorporated places. The annual P-26 series
includes revised estimates for all levels except incorporated places.
The estimation methodology is unique for each State, but generally
includes an average of three methodologies, with an appropriate “rake” to
insure that areas sum to the Statewide total.
• Regression (ratio-correlation) Method. A multiple regression
equation is used to relate changes in a number of different data series
to change in population distribution. Independent variables may
include automobile registrations, elementary school enrollments,
resident births for various periods, and Federal income tax returns,
among others.
• Component Method II. This method employs vital statistics to
measure natural increase and school enrollment to measure net
migration. These estimates are specific to the civilian population
under age 65. To this are added an estimate of the population 65 and
older based on Medicare statistics and an estimate of the resident
military population based on Department of Defense data.
• Administrative Records Method. This is an alternative component
method which uses individual Federal income tax returns to measure
civilian inter-county migration, and reported birth and death
statistics to estimate natural increase.

5-10 LAUS Program Manual

Statewide

Estimates

6

Development of Statewide
Estimates

Background
Historically, CPS samples have not been
sufficiently large, in all but the most
...an econometric approach
populous States, to produce reliable
to estimate State labor
force data.
monthly estimates directly from the
survey. 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 1970’s, a
six-month moving average ratio adjustment to CPS levels augmented the
Handbook estimate. In the late 1970’s, 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.
The Levitan Commission

Building on work done by Mathematica Policy Research under contract to BLS,
preliminary models were developed in the mid-1980’s. 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. Regression and time series
techniques were employed, with the models extensively evaluated using empirical
methods as well as recognized statistical theory.

LAUS Program Manual 6-1

Background

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 the State’s data. The changing coefficients are
estimated by the Kalman Filter, a widely used statistical technique that
evaluates structural change against sampling variability. The Kalman
Filter thus 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.
During the early 1990’s, 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 Kalman Filter. The Signal-Plus-Noise
model was implemented in January 1994. In 1996, time series modeling
was extended to the 11 more populous states because of reductions in the
size of the CPS sample.

6-2 LAUS Program Manual

Signal-Plus-Noise Estimation Model

Signal-Plus-Noise Estimation Model
The Signal/Noise Approach
The estimates of CPS data used by the LAUS
program consist of a true employment or
unemployment level, or “signal,” in addition to
a certain amount of “noise.” That is, 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 the
entire population.
CPSt = Signalt + Noiset

(EQ 1)

where:
CPSt = CPS employment or unemployment at time t.
Signalt = True employment or unemployment value at time t.
Noiset = CPS sampling error at time t, outliers, and irregular movements.
The signal/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 are known, or at least 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 noise. The goal of
the models is to isolate the signal from the noise to avoid distortions in the CPS
estimates and obtain the best possible estimates of the true labor force values.
The models for the employment and unemployment estimates are a combination
of two processes: a signal estimation and a noise estimation. The signal is a time
series model that is based on historical data relationships that are used in
estimating current true labor force values, 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 (implemented by the
Kalman Filter) that automatically adjusts the regression coefficients and trend and
seasonal components to adapt to gradual structural changes as they occur.

LAUS Program Manual 6-3

Signal-Plus-Noise Estimation Model

Sudden, unpredictable changes in the time series relationships are
handled by incorporating outlier effects into the model.
The noise estimation clears up the distortion caused by the CPS sampling
error and accounts for outliers and irregular movements in the data.

The Signal and Noise formulas are:
• Signal = Variable Coefficient Regression Component + Trend +
Seasonal + Signal Outliers
• Noise = CPS sampling error + Noise Outliers + Irregular
The signal estimation process consists of Regression, Trend, and Seasonal
components. The Regression component allows the coefficients of the
explanatory variable to vary over time. The Trend and Seasonal
components add even more flexibility to the models by accounting for
trend and seasonal variation in the CPS not explained by the regression
component. The noise estimation process controls for CPS sampling error
and historical outliers, and has a component to account for irregular
movements as well. The employment and unemployment rate model
specifications, as well a technical description of signal/noise models, are
included later in this chapter.

Accounting for CPS Sampling Error
There are two properties of the CPS, all controlled through the models,
that affect the time series data: changing reliability and the 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 historical data and current
CPS estimates corrected by a model-based estimate of sampling error.
The 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.

6-4 LAUS Program Manual

Signal-Plus-Noise Estimation Model

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 (prediction) + wt (CPSt - Nt)

(EQ 2)

where:
Signalt = the model estimate of the signal.
wt = the weight, between 0 and 1, given to the current CPS.
Signalt (prediction) = the model-based prediction of the signal based on
historical relationships.
Nt = the noise.
Nt (prediction) = the prediction of the noise.
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, one-eighth 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

LAUS Program Manual 6-5

Signal-Plus-Noise Estimation Model

% Identical HHs

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.

Overlap of Identical Households
80 75
60
50
50
38
38
40
25
25
25
13
13
20
0 0 0 0 0
0
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

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 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 standard error estimates 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

6-6 LAUS Program Manual

Signal-Plus-Noise Estimation Model

on 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 time profile 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.

Time Profile of CPS Error
1.2
1

Weights

0.8

Employment
Unemployment

0.6
0.4
0.2
0

0

2

4

6

8

1

1

1

1

1

2

2

2

2

2

3

3

3

3

3

4

4

4

4

4

5

5

5

5

5

6

6

6

0

2

4

6

8

0

2

4

6

8

0

2

4

6

8

0

2

4

6

8

0

2

4

6

8

0

2

4

Number of Months Ahead

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. The following table
quantifies the effects that autocorrelated errors have on variation in the CPS for a
typical State. This table shows the percent contribution of sampling error to the
CPS over increasing spans of time. For employment, about 57 percent of the

LAUS Program Manual 6-7

Signal-Plus-Noise Estimation Model

variation in the CPS from month to month is accounted for by sampling
error. If this error were completely random, its contribution to over-theyear variation in the CPS should be small, since normally the trend-cycle
movements in the signal dominate over that span of time. However, the
table indicates that this does not happen—sampling error accounts for 58
percent of the over-the-year variation in the CPS.

While not as strong as for employment, the long-run effect of the error on
unemployment is still very important.
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.
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—a nonrepresentative
sample resulting in a noise outlier, or a real non-repeatable event, such as
bad weather, strikes, etc., that cause an outlier in the signal. 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.

6-8 LAUS Program Manual

Signal-Plus-Noise Estimation Model

The purpose of the model is to capture the normal time series behavior of the
signal. Thus, the model must be flexible enough to adapt to structural changes in
the signal, but if 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 time series data:
1.) An additive outlier (AO) affects the series for only one month, such as a
sudden increase followed by a decrease.
2.) 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
3.) 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 and implemented during the annual
reestimation of the models.

LAUS Program Manual 6-9

Description of the Employment Model

Description of the Employment Model
Overview
The basic form for the LAUS signal/noise employment model is a
regression equation that uses the monthly CPS employment-to-population
ratio as the endogenous variable and the CES employment-to-population
ratio as the explanatory variable. Each State employment model can be
thought of as having a regression equation form, with a variable
coefficient component (CES employment-to-population ratio for the
employment model) and two time series components, which reflect the
State's CPS seasonal and trend movements not accounted for by the CES.
Chapter 4 discusses the conceptual and coverage differences between the
CPS and CES series. While the time series model adjusts for these
differences, knowledge of the CPS/CES differences is important to
understanding the nature of the model's CES variable. The model
accounts for these differences automatically because the regression
coefficients, residual trend and seasonal components are computed
separately for each State using State-specific data. Knowledge of the
survey differences is also useful in understanding why the trends in these
two series diverge at times.
Illustrated below is the basic structure of the employment model. The
model consists of three components.
E/P = EP Signal + EP Noise

(EQ 3)

Signal = bCESEP + Trend Residual + Seasonal Residual
Noise = CPS Error, Irregular, Transitory Outliers

Description of Signal Component
CES Base Variable

From the CES survey, a monthly estimate is developed of the total number
of persons on establishment payrolls who received pay for any part of the
employer's pay period that includes the 12th of the month. In the model,
the CES employment-to-population ratio is used as the major data source
for the model's target: the employed portion of the labor force. Data for
major strikes are added to the CES estimate from which the employmentto-population ratio is calculated.

6-10 LAUS Program Manual

Description of the Employment Model

Time Series Components

The part of the signal that is unaccounted for by the CES variable is represented
by the residual seasonal and trend components. The trend component is adjusting
for long-run systematic differences between the CES and the CPS series. Time
trend equations with variable intercepts and slopes are used to estimate the trend.
By allowing the parameters of the trend component to change, the estimated trend
component can adapt to change in the data. If there are frequent changes in the
level and/or slope of the trend, more weight is given to recent observations in
estimating the trend.
The seasonal component is estimated by fitting six trigonometric functions of time
with periodicities no longer than 12 months in duration to the data. As is the case
for the trend, the component is allowed to vary to permit adaptation to changing
relationships between the seasonal patterns of the CPS and CES. Because of
definitional differences between the CPS and CES, this component is necessary.
Differences in seasonality between the two series occur principally because there
are large seasonal variations in employed persons on unpaid absences who are
counted as employed in the CPS but not in the CES, and to seasonal variation in
agricultural employment. For a complete discussion of these differences, see
Chapter 4.
See Table A, page 6-23, for employment-to-population model specifications.

LAUS Program Manual 6-11

Description of the Unemployment Rate Model

Description of the Unemployment Rate Model
Overview
The basic form for the LAUS signal/noise unemployment model is a
regression equation that uses the monthly unemployment rate as the
endogenous variable and the unemployment insurance claims rate as the
explanatory variable. Each State model can be thought of as having a
regression equation form, with a variable coefficient component and two
time series components which provide flexibility for the State's CPS
seasonal and trend movements not accounted for by the explanatory
variable.
Below is the mathematical representation of the basic structure of the
unemployment rate model.
The model consists of three variables.
Urate = Urate Signal + Urate Noise

(EQ 4)

Urate Signal = bCLRST + Trend Residual + Seasonal Residual
Urate Noise = CPS Error, Irregular, Transitory Outlier

Description of Signal Component
The Base Variable—UI Claims Rate

The most important variable in the unemployment rate model is the UI
claims rate. This rate is defined in percentage terms as the ratio of State
continued claimants without earnings to total State CES employment.
This is a relative measure of the number of workers who are currently
unemployed and receiving UI benefits. Since the CES data are adjusted to
include strikers (CESADJ), the continued claimant count should exclude
any known strikers. (In some States, strikers may be eligible for
unemployment compensation). The statewide estimate of continued
claimants without earnings follows the standards outlined in Chapter 3.
The main weakness of the claims data is that these data are the by-product
of the UI tax system and therefore are subject to changes in the State’s
laws and administration, making them a biased cyclical indicator. Since
1980 there has been a marked deterioration in the cyclical sensitivity of
the claims data. Several factors may account for this. For example, in the
latter stages of a severe recession, many workers exhaust their UI benefits
and are dropped from the administrative count. As a result, the UI

6-12 LAUS Program Manual

Description of the Unemployment Rate Model

claimant count tends to diverge from the total rate when unemployment is high
and converge with the total rate when unemployment is low. Nationally, there have
been changes in the long-term relationships between the total unemployment rate
and the claims rate. From the 1970’s through the 1980’s, the proportion of
unemployed job losers collecting UI benefits nationally dropped from 75 percent
to less than 50 percent. This has reduced the sensitivity of the claims rate to
recessions. The seasonal pattern of the claims data also differs in important ways
from the CPS. Most notably, in the summer months, the entry of students into the
labor force is not reflected in the UI data. The seasonal component in the model
controls for this as well as for other seasonal differences. The model controls for
the cyclical bias in the UI data by changing the magnitude of the regression and
trend components. (See Chapter 3 for a discussion of the differences between the
CPS and UI systems.)
Time Series Components

The part of the signal that is unaccounted for by the claims rate is accounted for by
the seasonal and trend variables. Time trend equations with variable intercepts and
slopes are used to estimate the trend. By allowing the parameters of the trend
component to change, the estimated trend component can adapt to change in the
data. If there are frequent changes in the level and/or slope of the trend, more
weight is given to recent observations in estimating the trend.
The seasonal component is estimated by fitting six trigonometric functions of time
with periodicities no longer than 12 months in duration. As is the case for the
trend, the coefficients are allowed to vary to permit adaptation to changing
seasonal patterns.
See Table B, page 6-24, for unemployment rate model specifications.

LAUS Program Manual 6-13

Detailed Description of the Estimation Process

Detailed Description of the Estimation Process
Time Series Parameter Estimation
The models are made up of two parts: the signal and the noise. These
parts, jointly estimated, sum to the observed CPS estimate. The CPS labor
force estimate can be represented as follows:
y(t) = θ(t) + e(t)

(EQ 5)

where:
y(t) = the observed CPS
θ(t) = signal

e(t) = noise
The signal and noise portions of this equation can be broken down even
further. The signal is decomposed into a regressor component, a trend
component, a seasonal component, and an irregular component.
The form of the equation for the signal estimate is:
θ(t) = M(t) + T(t) + S(t) + I(t)

(EQ 6)

where the terms on the right side of the equation denote the regressor,
trend, seasonal, and irregular respectively.
Description of Signal Components

The regressor component is comprised of an observable economic
variable (either the CES or the UI) and a coefficient. The coefficient can
be variable or fixed. If the coefficient is variable, it is modeled as a
“random walk.” This means that the coefficient is equal to its previous
period’s value plus a current random disturbance.
M(t) = B(t) X(t)

(EQ 7)

B(t) = B(t-1) + VB(t)

(EQ 8)

If the variance of the random disturbance is zero, the coefficient will be
constant over time.
The trend component T(t), represented by the two equations below, is a
local approximation of a linear trend.

6-14 LAUS Program Manual

Detailed Description of the Estimation Process

When an approximation is local, it can allow for a changing trend over time. The
trend level is shifted by a variable VT(t) and the trend growth rate is shifted by the
white noise VR(t) of the variable. The trend component can be represented by the
following equations.
T(t) = T(t-1) + R(t-1) + VT(t)

(EQ 9)

R(t) = R(t-1) + vt(t)

(EQ 10)

If R(t) is zero, the trend follows a simple random walk in levels. If the variances
of the disturbances to the trend line are zero, the trend take a global linear form
with the intercept and slope fixed over the entire observation period.
The seasonal component is the sum of six trigonometric terms each within a
periodicity that repeats itself from 1 to 6 times within one year period. Each of
the six terms is subject to a white noise shock, Vj(t). Over a 12-month period, the
expected seasonal effects add to zero. If the variance of the seasonal component is
positive, the seasonal pattern is evolving over time. If the variance is zero, there is
a fixed seasonal pattern.
The irregular component, assumed to be stationary, is a residual not explained by
the trend or seasonal components. Sometimes the irregular component is just
white noise. If it is not white noise, it can be accounted for by using a low order
autoregressive (AR) process.
Description of Noise Term

The second part of the observed CPS estimate, the noise, represents CPS sampling
error. Sampling error can be accounted for through the variance-covariance
structure of the error term. Sampling error can take the form of both
autocorrelated error (caused by the 4-8-4 rotation scheme) and heteroscedasticity,
caused by sample redesigns, sample size changes, and a changing signal.
The error term can be represented as follows:
e(t) = λ(t)e*/e*(t)

(EQ 11)

The λ(t) variable represents the changing variance over time. This can also be
referred to as the variance inflation factor. The e(t) variable represents the
autocovariance structure which is assumed to follow an Auto Regressive Moving
Average (ARMA) process.
Description of Estimation Process

State agency staff prepare their official current monthly labor force estimates
using software developed at BLS that implements the Kalman Filter. This
algorithm is particularly well suited for the preparation of current estimates as

LAUS Program Manual 6-15

Detailed Description of the Estimation Process

new CPS data become available each month Since it is a recursive data
processing algorithm, it does not require all previous data to be kept in
storage and reprocessed every time a new CPS observation becomes
available. All that is required are estimates of the signal and noise
components and their variances for the previous month. This information
is combined with the current CPS data to produce an estimate of the
signal for the current month.

The Kalman Filter
Current estimates are calculated by the Kalman filter in a two-step
process.
Step 1.

Prediction

Make predictions of the signal and noise components for the current
month based on data up to the previous month. That is, predictions are
made for each of the components of the signal—regression, trend, and
seasonal—and the noise component (primarily sampling error).
Make a prediction of the CPS value for the current month by adding up
the predictions of the signal and noise components.
predicted CPSt = predicted Signalt + predicted Noiset
These predictions are the model based estimates made independently of
the current month’s CPS value.
Compute the variance of the error in predicting the CPS.
Variance of error in predicting CPSt = variance of error in predicting
Signalt + variance of error in predicting Noiset
Step 2.

Update

Use the current month’s CPS value to compute the model’s error in
predicting the current CPS value.
prediction errort = CPSt - (predicted Signalt + predicted Noiset)
= error in predicting Signalt + error in predicting Noiset
Update the predictions of the Signal and Noise components by allocating
the prediction error to the signal and noise components in proportion to
their relative accuracies.

6-16 LAUS Program Manual

Detailed Description of the Estimation Process

updated Signalt = predicted Signalt + gst * prediction errort
updated Noiset = predicted Noiset + gnt * prediction errort
gst = variance of error in predicting signalt/variance of error in predicting CPSt
gnt = variance of error in predicting noiset/variance of error in predicting CPSt
Note: gst + gnt = 1. Updated Signalt + updated Noiset = CPS t.
The quantities gst and gnt are referred to as the gain of the signal and the noise
component, respectively. To the extent that sampling error accounts for most of
the variation in the monthly CPS estimates, the gain for the estimated signal will
be low relative to the gain of the estimated noise component. In this way much of
the sampling error in the CPS is allocated to the noise component rather than to
the estimated signal.
Each month as new CPS data become available, the two basic steps of the Kalman
filter algorithm are repeated. That is, when new data become available at time
t
+ 1, the KF produces an updated estimate that incorporates the CPS data at t + 1,
but does not revise the previous period’s estimate of the signal and noise at time t
using the newly acquired CPS data. For this reason, the algorithm is referred to as
a “forward filter” since it always proceeds forward in time processing the latest
available data but never looks backward to revise earlier estimates with the most
current data. At the end of the year, all estimates are revised to incorporate all the
CPS data using the Kalman smoother.

Data Revision
Revised Estimates

In most statistical programs, revisions to estimates are made on a regular basis.
The schedule for these revisions depends on a variety of different factors related to
the availability of revised inputs, the amount of work involved, and the publication
schedule. In the LAUS program, estimates have usually been revised in two ways.
The first revision occurs one month after the initial (preliminary) estimate is
made. The primary reason for this “prior month” revision is that better, more
complete, input data are generally available by the time the following month’s
estimate is to be made. The second type of revision to the LAUS data occurs
during the annual benchmarking process. See Chapter 10 for a more detailed
discussion about annual processing.
Regression estimation has provided an additional source of information for
revising estimates. In time series regression, data from one month becomes input
for estimating subsequent months. This is true of both fixed and variable

LAUS Program Manual 6-17

Detailed Description of the Estimation Process

regression models because the time series estimates have strong
relationships over time, and each month’s estimate is a potential benefit to
all future months' estimations. The result is that the best estimate is
obtained when all the available data are used in a monthly estimate. The
process of using all the available data to estimate a given month is called
“smoothing,” and this technique is used in the annual reestimation and
benchmarking process.

6-18 LAUS Program Manual

Seasonal Adjustment of Statewide Estimates

Seasonal Adjustment of Statewide Estimates
Seasonality is defined as repetitive behavior that occurs on a regular calendar
basis, i.e., in cycles with periodicity that is annual, semiannual, quarterly,
monthly, or some other calendar unit. Seasonal fluctuations occur in labor force,
employment and unemployment levels. Examples of seasonal events include
weather changes, school schedules, industrial production schedules, harvests, and
major holidays. Fluctuations occur in the form of peaks and troughs and can
sometimes be observed simply by direct observation of a time series. However,
seasonal fluctuations often may not be immediately distinguishable from other
fluctuations in the data. Seasonal adjustment attempts to remove the seasonal
fluctuations in a time series. Once the seasonality is removed, the resulting series
can aid in further estimation and in the observation of underlying trend and
cyclical movements.
Research into the feasibility of seasonally adjusting LAUS model estimates began
with the formation of a workgroup in 1989. The workgroup evaluated the
appropriateness of applying X-11 ARIMA seasonal adjustment methodology to
model-based labor force estimates. The results of the testing indicated that
seasonal adjustment of the model-based series developed for this project was
satisfactory. X-11 ARIMA was able to remove seasonal variation in the modeled
series effectively without introducing distortions in the nonseasonal components
of the series. Adjustment factors related to the model series were similar to the
factors of the published full-sample CPS series, indicating that the models were
not forcing artificial patterns but rather were introducing seasonal patterns
consistent with the underlying CPS series.
Seasonal adjustment of State labor force estimates was introduced in 1992. Twice
each year, in July and January, seasonal adjustment factors for the labor force
series are produced for the next six months, updated to incorporate the experience
of the previous six months. Historical factors for the most recent five years are
updated during the January production of seasonal adjustment factors.

The Importance of Seasonal Adjustment
There are several reasons why seasonal adjustment is important. Without seasonal
adjustment, month-to-month comparisons of data within a time series are often
misleading. As a consequence, it is difficult to produce effective and meaningful
analysis over time without the availability of seasonally-adjusted series.
Seasonal adjustment is particularly important for State labor force statistics
because the data are used extensively to make economic decisions. Without
seasonally adjusted series, analysts, policy makers, and other data users would be

LAUS Program Manual 6-19

Seasonal Adjustment of Statewide Estimates

hindered in their efforts to make adequate assessments of economic
health, and they would not be able to accurately follow trends in State
labor markets.
The Seasonal Adjustment Process

Seasonal adjustment of time series data is designed to remove seasonal
fluctuations in order to better expose the trend or cycle. A seasonallyadjusted series is produced through the application of seasonal adjustment
factors to an original (not seasonally adjusted) series. These seasonal
factors are produced using the X-11 ARIMA seasonal adjustment
software. X-11 ARIMA is a model-based signal adjustment package that
allows the user to select the most appropriate options to obtain the best
seasonal adjustment results. Diagnostics and quality control information
provided as output from the application of X-11 ARIMA are used in the
option selection process.
In X-11-based procedures, seasonal adjustment can be additive (sum of
components equals original series) or multiplicative (product of
components equals original series). The additive model can be expressed
as follows:
O = TC + S + I

(EQ 12)

where:
O = original series.
TC = trend/cycle component.
S = seasonal component.
I = regular component.
Similarly, the multiplicative model can be expressed as follows:
O = TC*S*I

(EQ 13)

The basic step in X-11 seasonal adjustment is described below. The X-11
ARIMA expands upon these basic steps to produce reliable factors.
• Use a centered 12-month moving average of the data to extract the
trend from the original series to obtain a preliminary estimate of the
seasonal irregular.
• Estimate the seasonal component (factor). This is done by averaging
all values for the same month (stable seasonal) or by computing a
moving average of values for the same month.

6-20 LAUS Program Manual

Seasonal Adjustment of Statewide Estimates

• In additive terms: estimate S from S + I.
• In multiplicative terms: estimate S from S*I.
• Extract the seasonal components (factors) computed from the original series to
get the seasonally adjusted series.
• In additive terms: TC + I = O - S = adjusted series.
• In multiplicative terms: TC*I = O/S = adjusted series.
Note that in the last step, an additive seasonal factor is subtracted from the original
series and a multiplicative seasonal factor is divided into the original series. The
seasonal series may be further smoothed with a Henderson moving average and a
second set of seasonal adjustment coefficients may be produced. Several iterations
of the process may be necessary.

LAUS Program Manual 6-21

Seasonal Adjustment of Statewide Estimates

6-22 LAUS Program Manual

Seasonal Adjustment of Statewide Estimates

Table 6-1 Employment-to-Population Ratio Model Specifications, 1997
States

AL
AK
AZ
AR
CA
LA-LB
Balance
CO
CT
DE
DC
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
MI
MN
MS
MO
MT
NE
NV
NH
NJ
NM
NY
NYC
Balance
NC
ND
OH
OK
OR
PA
RI
SC
SD
TN
TX
UT
VT
VA
WA
WV
WI
WY

Regression
Variable

Trend Level Shifts

CESEP
CESEP
CESEP

Jan-94

CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP

Sep-90
Jun-84

CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP
CESEP

Jan-94
Feb-89

Jun-90
Dec-80
Feb-79
Oct-82

Sep-90

Std of coeff Std of Trend
Std of
Regression
change over change over Seasonal
coefficient
1 month
1 month change over
12/96
1 year span
span
span

Additive Outliers

Sep-81
Jun-85
Oct-84

Mar-91
Jan-83
Aug-87

Dec-88

Sep-87

Jun-90

Aug-81
Jan-80
Sep-78

Aug-82
Nov-87

Jul-85
Jul-85
Nov-84

Nov-90

Jan-92
Aug-90

Apr-81

Oct-82
Jul-84

May-86

Apr-85
Nov-90

Jan-87
Dec-90

Feb-87

May-87

Jan-83
Jul-83

May-87
Dec-91

Apr-89

Feb-91

Aug-90

Jan-94

Dec-90

Nov-84

Apr-85

Jun-84
Jan-90
Sep-85
Nov-93

Nov-80
Aug-81

Oct-83

Dec-79
Jan-89
Sep-85
Feb-80
Dec-85
Jun-88

Jan-80

Jan-83
Sep-88

Oct-82
Nov-83

Dec-91
Mar-80

Jan-91

0.686
.
0.401
0.518

0
.
0
0

0
0
0.1397
0

0.0063
0.0277
0.0116
0.0143

0.424
0.315
0.567
0.442
0.432
0.211
0.69
0.459
0.275
0.426
0.562
0.667
0.504
0.405
0.492
0.605
0.503
0.668
0.562
0.858
0.568
0.513
0.832
0.645
0.58
0.335
0.559
0.787
0.675

0
0
0
0
0.0008
0
0
0
0.0009
0.001
0.001
0.0015
0.0019
0
0.0001
0.0025
0.0014
0
0
0.0001
0
0
0
0.0004
0
0.0005
0
0
0

0.204
0.1398
0.0579
0.0246
0
0.2155
0.0236
0
0
0.1699
0.1366
0
0
0
0.0258
0
0
0.0112
0.0056
0.1378
0.0106
0.0176
0
0.0372
0
0
0
0.0199
0.0202

0.0305
0
0.0112
0
0.0292
0.0038
0
0.0315
0
0.012
0.0001
0
0.0162
0
0.0004
0.0066
0
0.0096
0
0.0104
0.0015
0
0.0147
0.0047
0.0137
0.0085
0
0
0.0035

0.647
0.857
0.683
0.172
0.729
0.479
0.345
0.669
0.609
0.359
0.354
0.651
0.519
0.437
0.819
0.375
0.828
0.783
0.814
0.409

0.0004
0.0004
0
0
0
0
0.0044
0
0.001
0
0
0
0
0.0004
0
0
0
0
0
0

0
0.0961
0.1085
0.1883
0
0.052
0
0.0415
0.0079
0.0191
0
0
0.0218
0
0
0
0.0697
0
0
0.0013

0
0.0003
0.0312
0.0061
0
0.0029
0
0.0059
0.0219
0.0154
0
0.0052
0
0.0032
0.0541
0
0.0333
0
0
0

NOTE: A measure of coefficient stability is provided by the standard deviation (Std) of the change in the coefficient over a one month span. An
exact value of zero for the standard deviation indicates the respective (smoothed) component has a constant value over time. For example,
a Std of zero for the CESEP regression coefficient in AZ's EP model means that the coefficient is fixed at a value of 0.401 for the historical
series 1976-96.

LAUS Program Manual 6-23

Seasonal Adjustment of Statewide Estimates

Table 6-2 Unemployment Rate Model Specifications, 1997
States

AL
AK
AZ
AR
CA
LA-LB
Balance
CO
CT
DE
DC
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
MI
MN
MS
MO
MT
NE
NV
NH
NJ
NM
NY
NYC
Balance
NC
ND
OH
OK
OR
PA
RI
SC
SD
TN
TX
UT
VT
VA
WA
WV
WI
WY

Regression
Variable

Trend Level Shifts

CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRSTFE
CLRST
CLRSTFE
CLRST
CLRFADJ
CLRST
CLRST
CLRST
CLRSTFE
CLRSTFE
CLRSTFE
CLRADJ
CLRSTFE
CLRST
CLRSTFE
CLRADJ
CLRSTFE
CLRST
CLRSTFE
CLRST
CLRST
CLRSTFE
CLRST
CLRADJ

CLRSTFE
CLRST
CLRSTFE
CLRSTFE
CLRADJ
HBEXR
CLRFADJ
CLRSTFE
CLRSTFE
CLRSTFE
CLRST
CLRSTFE
CLRSTFE
CLRST
CLRSTFE
CLRSTFE
CLRST
CLRSTFE
CLRST
CLRST

Std of coeff Std of Trend
Std of
Regression
change over change over Seasonal
coefficient
1 month
1 month change over
12/96
span
span
1 year span

Additive Outliers

Jul-77

Sep-90

Mar-91

Jan-92

Aug-92

Nov-93

Jul-82

Aug-92

Jul-84
May-80

Jan-88
Oct-80

Oct-89
Sep-83

Apr-88

Apr-92

Oct-78

Mar-81

Mar-79
Sep-84

Nov-81
Jan-94

Jul-78
Jun-82
Dec-77
Mar-85
Aug-90
Jan-79
Apr-79
Aug-77

Aug-82

May-91

May-83

Feb-90

Oct-84

Aug-80
Aug-79

Jul-84

Apr-84

Mar-90

Jun-80

Apr-80

Dec-81

Jun-84

May-92

Jun-84
Jun-77
Oct-81

Apr-89
Jul-83

Jun-80

Dec-80

Jan-88
Mar-83
Feb-79
Jul-81

Sep-86
Dec-89

Jan-79

Apr-87

Sep-88
Dec-89

Feb-86

Jan-94

Mar-90
Jun-87
Jan-88
Dec-82
Jun-84

Jan-94
Apr-83

Jan-90

Jun-87

Apr-81

Jul-91

Jan-94

Jun-91

May-86

1.47
0.86
2.029
0.756

0.0081
0.0056
0.0165
0.0063

0.0759
0.0286
0.0381
0.0323

0
0.0119
0.0063
0.0374

1.89
0.836
2.133
1.317
0.43
1.609
1.177
1.339
1.107
1.076
1.157
1.004
1.027
1.007
0.987
0.662
1.037
1.328
1.25
0.561
1.136
1.014
0.449
0.658
1.363
1.61
0.766
1.261
2.006

0.0284
0.0694
0.0079
0.0011
0
0
0.024
0
0
0
0.0071
0.0007
0
0
0.0244
0
0
0.0049
0
0
0
0.0051
0.0032
0
0.0058
0
0
0.0129
0.001

0.0246
0.0615
0.0202
0.0239
0.1672
0.1666
0.0005
0.2089
0.0535
0.0351
0.1513
0.1071
0.0277
0.0877
0.0076
0.1385
0.0526
0.0203
0.0948
0.2505
0.0419
0.1439
0.0888
0.0678
0.0252
0.0589
0.102
0.1447
0

0
0.0004
0.0136
0.0112
0.0011
0
0.0078
0
0.0087

0
0.0102
0
0.0065
0
0
0.0051
0
0
0.0154
0
0
0.0086
0
0
0.0041
0

0.57
0.996
1.457
1.141

0
0
0.0133
0.0014

0.2724
0.1536
0.1566
0.0127
0.1225

0
0
0
0
0.0018

1.373
0.862
0.562
0.965
1.463
1.441
1.461
1.372
1.362
1.753
1.029
0.96
0.876
1.672

0.001
0.0148
0.007
0
0
0.0113
0.0086
0
0.0195
0
0.0111
0
0.0159
0.004
0.0087

0.1992
0.0328
0.0452
0.157
0.1084
0.0119
0.0507
0.0877
0.0129
0.0016
0.0065
0.0555
0.1055
0.0752
0.0122

0
0
0.0201
0.0081
0.0241
0.015
0.0033
0
0.0084
0.0069
0
0.0146

0.0068

0
0.0143

NOTE: A measure of coefficient stability is provided by the standard deviation (Std) of the change in the coefficient over a one month span. of
An exact value zero for the standard deviation indicates the respective (smoothed) component has a constant value over time.
For example, a Std of zero for GA's residual seasonal component means that the monthly seasonal factors, while they differ
from month-to-month, are fixed from year-to-year.

6-24 LAUS Program Manual

Labor Market

Estimates

7

LAUS Estimation: Labor Market
Area Estimates

Introduction
n the late 1940’s when subnational 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. Labor market areas (LMAs) are 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. Labor market areas are defined in terms of full counties
in all areas except New England, where Minor Civil Divisions (MCDs) are used to
define LMAs. In the criteria below, the term “county” includes county equivalents
and, in New England, refers to MCDs.

I

A general definition for a labor market area is 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 (MAs) or small labor
market areas, and exhaust the geography of each of the States, the District of
Columbia, and Puerto Rico. The Office of Management and Budget (OMB) is
responsible for defining the MAs while the Local Area Unemployment Statistics
Division (LAUS) of the BLS performs this function for small labor market areas.
Currently, there are 332 metropolitan areas and 2,049 small labor market areas.
Metropolitan areas are designated on the basis of population, urban area, and
commutation data. Since population and urban area data are inappropriate for
defining the generally less populous small labor market areas, commutation data
are used to determine which counties are deemed single-county areas and which

LAUS Program Manual 7-1

Introduction

are combined into multi-county areas. Regardless of population size,
commuting flows are an indication of the degree of integration of labor
markets among counties; commutation data show the extent to which
workers have been willing and able to commute to other counties.
Every 10 years the Nation’s system of labor market areas is reevaluated
and redefined, using the latest Decennial Census information on
population and commutation.

Standards for Defining Metropolitan Areas
The general concept of a metropolitan area is that of a core area
containing a large population nucleus, together with adjacent
communities that have a high degree of economic and social integration
with that core. Included among metropolitan areas are Metropolitan
Statistical Areas (MSAs), Consolidated Metropolitan Statistical Areas
(CMSAs), and Primary Metropolitan Statistical Areas (PMSAs). In
addition, New England County Metropolitan areas (NECMAs) are an
alternative set of areas defined for the six New England States.
Metropolitan Statistical Area (MSA)

An MSA consists of one or more counties that contain a city of 50,000 or
more inhabitants, or contain a Census Bureau-defined urbanized area and
have a total population of at least 100,000 (75,000 in New England).
Counties containing the principal concentration of population—the
largest city and surrounding densely settled area—are components of the
MSA. Additional counties qualify to be included by meeting a specified
level of commuting to the counties containing the population
concentration and by meeting certain other requirements of metropolitan
character, such as a specified minimum population density or percentage
of the population that is urban. MSAs in New England are defined in
terms of cities and towns, following rules concerning commuting and
population density.

7-2 LAUS Program Manual

Introduction

Consolidated Metropolitan Statistical Area (CMSA)

A CMSA is a metropolitan area that has a population of at least 1 million and
which has been divided into two or more PMSAs. A CMSA comprises the same
geographical area as its constituent PMSAs.
Primary Metropolitan Statistical Area (PMSA)

Subareas may be defined within an area that meets the requirements to qualify as
an MSA and also has a population of one million or more. The definition of these
subareas, called PMSAs, requires meeting specified statistical criteria and having
the support of local opinion. A PMSA consists of a large urbanized county or
cluster of counties (cities and towns in New England) that demonstrate strong
internal economic and social links in addition to close ties with the central core of
the larger area. Upon the recognition of PMSAs, the entire area of which they are
parts becomes a CMSA. All territory within a CMSA is also within some PMSA.
New England County Metropolitan Areas

NECMSAs are county-based alternatives to the city-town based MSAs and
CMSAs in the six New England States. The county composition of a NECMA
reflects the geographic extent of the corresponding MSA(s) or CMSA(s).
NECMSAs are not defined for individual PMSAs.

Standards for Defining Small Labor Market Areas
The following criteria were applied to those areas of the Nation not already
included in the OMB-defined Metropolitan Areas, in identifying small labor
market areas:
1.) If 15.0 percent (rounded to the nearest tenth of one percent) or more of
the employed workers residing in a county commute to another county,
then the two counties were combined into one labor market area.
2.) After combining counties based on the “county-to-county” commuting
criteria in (1) above, county-to-LMA flows were checked. If two or more
counties were combined into one LMA, an additional county was added if
15.0 percent or more of the employed workers residing in that additional
county commuted to the combined LMA (not necessarily to any one
county within the LMA). This procedure may have required several
iterations.
3.) In the case of existing multi-county LMAs, attempts were made to
maintain historical continuity. Thus, if the commuting flows were only
marginally below the new criteria (that is, within 1.645 standard errors of
15.0 percent), the counties remained combined.

LAUS Program Manual 7-3

Introduction

4.) Counties were first combined based on the commutation criteria,
and then potential multi-county LMAs were checked for
contiguity. Noncontiguous portions of potential LMAs were
considered separately. If the noncontiguous area contained more
than one county, it was reevaluated using (1), (2), and (3) above.
If the noncontiguous area was a single county, it was designated
as a single LMA.
5.) In the New England States, due to the number of small cities and
towns, residual MCDs were added to contiguous multi-area LMAs
based on commuting flows and/or other economic ties. That is, if
after applying the commutation data, an MCD had been identified
as an individual LMA, the State could recommend that the MCD
be added to a contiguous multi-area LMA, especially if the MCD
was extremely small.

Handbook Estimates
Estimates for the LMAs, which
geographically exhaust each State, are
produced independently by a building
block approach which uses current
unemployment insurance 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 and
decennial census relationships. This
methodology is referred to as the Handbook procedure.
When the Handbook was first introduced as a standard procedure for
subnational 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 utilized a
system of estimates which was reflective of the employment and
unemployment structure in terms of coverage for UI in the 1960’s. Over
the years, refinements were made to the components and to the 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. In
addition, the Handbook procedure has been streamlined to reflect
expanded UI coverage and economic and behavioral changes in the labor
market.
Additivity of the substate estimates to the State estimates was introduced
in 1977, to address a methodology issue and a Federal program allocation
need. The sum of substate area unemployment was significantly less than
the State level because of greater difficulty in estimating new and

7-4 LAUS Program Manual

Introduction

reentrants to the labor market for areas. Forcing area unemployment estimates to
the State total corrected for this methodological problem in a proportional manner.
This correction also allowed for the complete, to-the-dollar distribution of federal
funds to areas when LAUS unemployment was used as the allocation algorithm.
(Additivity is discussed in detail in Chapter 8.)

LAUS Program Manual 7-5

Labor Market Area Employment

Labor Market 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, and 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 detail.)
In brief, total employment is composed of nonagricultural wage and
salary employment, all other nonagricultural employment, and
agricultural employment.

Nonagricultural Wage and Salary Employment
The nonfarm payroll estimate for a particular labor market area may be
based on data from several sources. The principal source is the estimate
from the CES survey for the area. If the LMA is not covered by the CES
program, a sample-based employment series developed under State
auspices is the next best source.
For small LMAs without sample-based employment estimates,
extrapolated quarterly ES-202 employment data are used. The current
quarter’s monthly employment estimates are projected based on actual
same quarter one year ago and previous quarter one year ago data. That is,
to determine an area’s employment estimate for a given month, the overthe-year change in employment for that area is multiplied by its
employment level for the same month one year earlier. Employment
levels for each area are estimated using the following formulas:

M1c = (AMEp ÷ AMEpy) x M1cy
M2c = (AMEp ÷ AMEpy) x M2cy
M3c = (AMEp ÷ AMEpy) x M3cy

Where:
M1 = first month employment

7-6 LAUS Program Manual

Labor Market Area Employment

M2 = second month employment
M3 = third month employment
AME = average monthly employment
c = current quarter
p = prior quarter
cy = current quarter one year ago
py = prior quarter one year ago
The “place-of-work” nonfarm employment estimates must be adjusted to a placeof-residence basis, as in the CPS. Estimated adjustment factors for several
categories of employment have been developed on the basis of employment
relationships which existed at the time of the most recent decennial census. These
factors are appropriately weighted and combined into a single factor which is then
applied to the place-of-work nonfarm employment estimates for the current period
to obtain nonfarm wage and salary employment estimates adjusted to place of
residence.
Nonagricultural wage & salary employment

+ Labor disputants

933,100
0

Unadjusted wage & salary employment

933,100

Unadjusted wage & salary employment

933,100

× Residency adjustment ratio
Total nonagricultural wage & salary employment

0.967100
902,401

“All-Other” Employment
The term “all-other” employment represents the self-employed, unpaid family,
and private household workers (domestics) in the nonagricultural industries.
These people work, but generally are not listed on payroll records. Two sources of
all-other employment data exist—the decennial census and the CPS. The census
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, CES estimates of nonagricultural wage and salary data, and census
counts of area nonagricultural wage and salary employment and all-other
employment data. The census counts are used as a base period estimate.

LAUS Program Manual 7-7

Labor Market Area Employment

This section explains the development of estimates of self-employed,
unpaid family, and private household employment. During intercensal
years, data for all-other employment are available in the CPS. While these
data are published on a monthly basis only for the nation as a whole,
unpublished data are available at the State level. Research has shown that
the national and State CPS data, together with national and area wageand-salary employment, can be used to extrapolate the census all-other
benchmark for areas.
The original analysis which led to the first estimating methodology was
based on an examination of the relationship between all-other
employment and wage-and-salary employment in the Nation as a whole
and in a randomly selected sample of areas using the 1940 and 1950
census data. It was found that, in both the areas and the U.S. total, the
relative change in wage-and-salary employment was accompanied by a
proportional relative change in all-other employment. In other words,
slow wage and salary growth was accompanied by slow all other
employment growth, and rapid wage and salary growth was accompanied
by rapid all-other growth.
It was also found that the proportional relative changes in all-other
employment in the areas and in the Nation were very close to each other.
This meant that the relative change in area all-other employment could be
derived given the relative change in area wage-and-salary employment
and the ratio of the relative national change in all-other employment to the
relative national change in wage-and-salary employment.
Analyses utilizing data from subsequent censuses corroborated the
findings of the original study. However, discrepancies between individual
areas, on the one hand, and areas and the Nation on the other, proved quite
common, and pointed out the need for area adjustment. The CPS sample
expansion of the 1970’s provided additional geographic detail on all-other
employment and allowed the opportunity for analysis and testing of
differences in the proportionality factor between States.
Following each census, each State’s k value, where k equals the relative
change in wage and salary employment divided by the relative change in
all-other employment, is calculated and reviewed. Clusters of States with
similar proportionality constants are grouped into strata. Four strata were
defined following the 1980 census; three were defined following the 1990
census. By grouping States into strata based on their ratio of relative
change, it was found that all-other employment estimates could be
improved. Specifically, using the strata ks for estimating the all-other
employment significantly reduced the range of error in estimating allother employment.

7-8 LAUS Program Manual

Labor Market Area Employment

An area k-value assigns areas to one of the three strata developed following the
1990 census. It remains in use for the entire intercensal period for the application
of a Step 3 ratio provided monthly by BLS.
For each area, the area k is calculated with the following formula:

w
w
k = a t ÷ a t-1
t ÷ t-1

Where:

k = proportionality constant for area data
w = area March/April nonagricultural wage and salary employment
a = area Census all-other employment
t = current time period: 1990 for Census all-other employment; March/April 1990
for nonagricultural wage and salary employment
t-1 = prior time period: 1980 for Census all-other employment; March/April 1980
for nonagricultural wage and salary employment
Strata from the 1990 Census are as follows:
Stratum 1

Stratum 2

Stratum 3

__________________________________________________
(k < 0.900)

(0.900< k < 0.990)

(k > 0.990)

Estimating All-Other Employment

Each month, BLS calculates a Step 3 Ratio for each stratum and provides the
Ratios to the States. States provide their current month’s statewide nonagricultural
wage and salary employment estimates to BLS according to a schedule provided
annually. The following steps are used by BLS to calculate the Step 3 Ratios.
Step 1.

Determine the all-other employment change ratio. The current
month CPS all-other employment estimates for each State in the
stratum are summed and divided by the sum of the March/April 1990
census all-other employment estimates for the States in the stratum.

LAUS Program Manual 7-9

Labor Market Area Employment

Step 2.

Determine the nonagricultural wage and salary
employment change ratio. The current month CES
nonagricultural wage and salary estimates for each State in the
stratum are summed and divided by the sum of the March/April
1990 (two-month average) nonagricultural wage and salary
estimates for the States in the stratum.

Step 3.

Determine the ratio of relative change for each stratum.
Divide each stratum all-other employment change ratio by the
corresponding stratum wage and salary employment change
ratio, i.e., Step 1 divided by Step 2.

Table 7-1 presents the monthly Step 3 ratios by stratum for 1990 through
September 1997
States utilize the Step 3 ratios to estimate all-other employment for each
LMA, as follows.
1.) Calculate the area nonagricultural wage and salary employment
change ratio. Divide the current month CES nonagricultural wage and
salary employment (including persons involved in labor-management
disputes) by the March/April 1990 nonagricultural wage and salary
employment.
Nonagricultural wage & salary employment

÷ March/April 1990 nonagricultural employment
Area employment change ratio

933,100
53,589
0.994663

2.) Calculate the area all-other employment change from the 1990
census to the current month. Multiply the ratio calculated above (the
nonagricultural wage and salary employment ratio) times the 1990 census
all-other employment estimate for the area.
Area Wage and Salary Employment Ratio

0.994663

× Census All Other Employment

53,589

All other employment change

53,303

7-10 LAUS Program Manual

Labor Market Area Employment

3.) Calculate the area all-other employment for the current month. Multiply
the estimate calculated above (the area all-other employment change figure) by
the stratum-specific Step 3 ratio provided by BLS.
All other employment change

× Step 3 ratio
Area All other employment

53,303
0.894000
47,653

Total Agricultural Employment
Due to the lack of an adequate current database, the
methodology for estimating agricultural employment utilizes
relationships between various series of agricultural
employment to project current agricultural employment. In
addition to the decennial census, the other sources of data used
for this purpose are the CPS and the Department of
Agriculture’s Agricultural Labor Survey (ALS). CPS data on
agricultural employment published in Employment and
Earnings are for the national level only, but the unpublished
CPS agricultural employment estimates for the States are indirectly incorporated
into the estimates as monthly change factors. The ALS data on agricultural
employment for States are used as part of the annual benchmarking process.
Neither the ALS nor the CPS yields current monthly estimates reliable enough to
set the level of farm jobs for the desired areas. From the ALS, agricultural
employment estimates for the first month of a calendar quarter are published for
fifteen regions and three States. The July level is used each year at benchmark
time to rebase to the decennial census level. From the CPS, unpublished monthly
agricultural employment estimates are available and are used to develop monthly
change factors by ALS region. (See Table 7-2 for monthly factors by ALS region
for 1990-1997.)
The ALS utilizes two sampling frames. The list frame of agricultural producers is
a stratified random sample of employers likely to have hired workers. The area
frame contains all land units in the nation and is used to compensate for the
incompleteness of the list of employers. Since the relative error on the number of
workers at the ALS regional level is smaller than the comparable relative error
using CPS data, it is used to adjust the agricultural employment estimate annually.
Unpublished CPS data, aggregated to ALS regions, are used to develop monthly
agricultural change factors.

LAUS Program Manual 7-11

Labor Market Area Employment

Defining the Census Benchmark

The agricultural employment benchmark level for all areas is established
by the decennial census. The 1990 agricultural data consist of workers in
the following industrial classifications:
Description

SIC

Agricultural production - crop

01

Agricultural production - livestock

02

Agricultural services

07 (part)

SICs 074 (Veterinary Services), 075 (Animal Services, except
Veterinary), and 078 (Landscape and Horticultural Services) are excluded
from the agricultural data. These industries are included in the CES
program in the monthly estimation of the Services Industry.
Forestry and fisheries (SICs 08 and 09) data may be included in the
census benchmark at the State’s option. States including these SICs
should notify the appropriate regional office. If a State chooses to include
data for these two industrial groups, estimates for all LMAs within the
State should also include such data.
Composition and Use of Agricultural Employment Estimating
Regions

The following table lists the States included in each of the agricultural
employment regions. All areas in each State generally use the factors
developed for the respective region as a whole. On an annual basis, the
July-to-July percentage change in the agricultural employment region is
used to adjust the census-based level for the LMA.

7-12 LAUS Program Manual

Labor Market Area Employment

Agricultural Employment Estimating Regions
Region

Geographic Region

State

1

Northeast I

CT, ME, MA, NH, NY,
RI, & VT

2

Northeast II

DE, MD, NJ, & PA

3

Appalachian I

NC & VA

4

Appalachian II

TN, & WV

5

Southeast

Al, GA, & SC

6

Florida

FL

7

Lake

MI, MN, & WI

8

Corn Belt I

IL, IN, KY, & OH

9

Corn Belt II

IA & MO

10

Delta I

LA, AR, & MS

11

Northern Plains

KS, NE, ND, & SD

12

Southern Plains

OK & TX

13

Mountain I

ID, MT, & WY

14

Mountain II

CO, NV, & UT

15

Mountain III

AZ & NM

16

Pacific

OR & WA

17

California

CA

18

Hawaii

HI

19

Michigan

MI

20

Minnesota

MN

21

Wisconsin

WI

The States have the option to make use of data from agricultural regions other than
their own. This may be done for any LMA in a State if local knowledge of the
agricultural economy indicates that another region better reflects local agricultural
employment. (In the case of interstate areas, the subarea calculation is determined
by the controlling State’s regional selection.) Once the selection of an alternate
regional factor has been made, it must be continued until the next census. The
production of all current and benchmarked data must also reflect this selection.
The use of alternative agricultural factors is requested using the atypical/exception
procedure.

LAUS Program Manual 7-13

Labor Market Area Employment

At benchmarking, the most recent July-to-July regional change factors are
made available for creation of the new July adjusted number. New
monthly change factors are also provided at this time to revise monthly
estimates from July of the previous year forward. The new July base is
then used until the next year’s benchmark. The preliminary AugustDecember estimates should therefore be replaced with revised estimates,
based on the new July adjusted base, providing a consistent series through
the next calendar year.
Description of Agricultural Estimating Procedure

Each year at benchmarking time, a new agricultural employment level is
set for the previous July with ALS data. Using this level and the similar
ALS level for the July one year earlier, the months from August through
June are wedged. The months from August following the most recent
ALS release through December are revised using the new July base level
and the CPS monthly change factors. (See Chapter 10.)
The current calculation is:

A= F × C
ij

ik

ij

where,
1.) A ij = total agricultural employment for month i in LMA j
2.) Fik = monthly factor for total agricultural employment for month i
in agricultural employment estimating region k, and
3.) C ij= 1990 decennial census agricultural employment benchmark
level for the LMA.

Census agricultural employment
× Monthly factor

Total agricultural employment

7-14 LAUS Program Manual

2,572
0.743000
1,911

Labor Market Area Unemployment

Labor Market Area Unemployment
In the current month, the estimate of unemployment is an aggregate of the
estimates for each of the two building-block categories, those unemployed
covered by UI and new and reentrants (the noncovered).

Covered Unemployment: Claims Data
Statistics from the UI systems are the only current measure of unemployment at
the substate level.
The “covered” category consists of (1) those who are currently collecting UI
benefits, and (2) those who have exhausted their benefits. Only the insured
unemployed are obtained directly from an actual count of current UI claimants for
the reference week under State UI, Federal, and Railroad programs. Two insured
unemployed counts, continued claimants and final payment recipients, are used in
the development of LMA unemployment estimates. Continued claimants are
persons certifying to a compensated or noncompensated week of unemployment
including the 12th of the month under the State UI and UCFE programs. (See
Chapter 3.)
The exhaustee component represents a significant part of the overall unemployed
estimates, second in size to the active insured count, 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 (and monthly
counts in some areas receiving intrastate commuter claims) 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.
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 addition,
persons receiving final payments in previous weeks or months are carried forward
in decreasing numbers into successive periods, by applying a CPS-based
“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.

LAUS Program Manual 7-15

Labor Market Area Unemployment

Based on research conducted by Hyman Kaitz, a formula which utilizes
the parallel relationship between the rate of unemployment and the
duration of unemployment spells, and quarterly average CPS State
duration data, yields survival rates for each of the four unemployment
ranges. On a monthly basis, areas can select a survival rate from these
four ranges which most closely relates to recent local unemployment rate
conditions. Thus, high unemployment rate areas select a higher survival
rate and have higher exhaustee levels and Handbook 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. Each quarter the 50 States and the
District of Columbia are divided into four unemployment rate groups.
Each of four ranges represents a set of States within a given range of
unemployment rates. Each group contains roughly 25 percent of the
quarterly average nationally-weighted unemployment.
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).
(See Table 7-3 for survival rates.)
2. Each month during a given quarter, areas select from among these
survival rates in developing LAUS estimates 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 twoquarter lagged relationship between the unemployment rate and the
survival rate and a one-quarter operational lag.
Calculation of Exhaustees from Weekly Final Payments

Following is an illustration of the steps involved in calculating exhaustees.
Each column of the worksheet is described below along with the
calculations needed to obtain a column’s result.

7-16 LAUS Program Manual

Labor Market Area Unemployment

Unemployed Exhaustee Worksheet
Est.
Month.

Month

1

2

7/95

4/96

LAUS
Rate

Weekly
Rate

Week
Ending

A.F.P.*

3

4

5

6

7

8

7.8

0.949

3/23

120

1,484

1,408

0.949

3/30

135

1,528

1,450

4/06

117

1,585

1,504
1,5380

P.W.*

U.E.*
(4 × 7)

4/96

7/95

4/96

7/95

-

0.949
0.949

4/13

87

1,621

0.958

4/20

100

1,652

1,557

0.958

4/27

72

1,657

1,587
1,589
1,589

4/96

7/95

-

5/96

8/95

8.1

5/96

8/95

5/96
5/96

8/95

-

0.958

5/04

93

1,659

8/95

-

0.958

5/11

114

1,659

LAUS Program Manual 7-17

Labor Market Area Unemployment

Column

Explanation / Entry

1

The month for which the LAUS estimate is made.

2

The month for use in survival rate selection.

3

Enter the State or area's LAUS unemployment
rate for the month indicated in Column 2.

4

Based on the Column 3 rate, enter the appropriate
weekly survival rate obtained from accessing
Sungard file:
&&&YBDBLS.A130.SURVIVAL:CHART

5

The week ending date from the reference week of
the last month of the prior quarter through the
current quarter.
Weekly Final Payment Data

6

AFP = Actual Final Payments. Final payments
collected on a weekly basis.

7

Sum of final payments for the previous week and
unemployed exhaustees from previous week.

8

Unemployed exhaustee estimate. Column 7 times
the Column 4 weekly rate.

Noncovered Unemployment
Unemployment for groups not covered by unemployment insurance is not
specifically calculated. Rather, noncovered unemployment is assumed to
be distributed proportionally among LMAs in relation to their share of
total unemployment, through the additivity process. This proportional
relationship may not be appropriate in States with a significant
agricultural component, particularly where UI coverage of agricultural
employment does not go beyond the Federal minimum. To address
situations where there is a relative concentration of agricultural
employment in small labor market areas, States can include a specific
agricultural unemployment component to the Handbook.

7-18 LAUS Program Manual

Labor Market Area Unemployment

Agricultural Unemployment
Direct estimation of agricultural unemployment may be used in States with at
least one LMA where agricultural employment is 25 percent or more of total
employment. States which qualify must obtain approval from BLS to estimate
agricultural unemployment directly. In such cases, this direct estimation must be
used in all labor market areas of the State. Other States may request approval for
atypical treatment of agricultural unemployment for a specific LMA if it can be
demonstrated that the lack of such an estimate has a deleterious effect on
estimates for that area.
In order to develop this estimate, the relationship between the unemployment rate
for agricultural workers and for those from nonagricultural activities is used. This
relationship varies monthly, reflecting differences in the seasonal patterns for the
two groups. Monthly fractional rates for estimating agricultural unemployment
are provided by the national office via the LAUS State software. These rates were
developed by combining separate rates for agricultural wage and salary and
agricultural self-employed and unpaid family workers. Each group was
appropriately weighted, based on CPS monthly employment levels for the
previous two years; the weighted rates are combined by addition. The appropriate
rate is then applied to the noncovered agricultural employment estimate.

Unemployed New Entrants and Reentrants
For many unemployed individuals, the current spell of unemployment has not
been immediately preceded by employment. These unemployed entered the labor
market from outside the labor force after having completed military service,
family responsibilities, education, or other situations. These individuals are known
as unemployed entrants.
Unemployed entrants can be further divided into two groups. One group includes
the individuals who enter the labor market for the first time and do not find jobs.
These persons are defined as unemployed new entrants. The other group includes
those who enter the labor market after period of retirement from the labor force
and are unable to find employment. They are defined as unemployed reentrants.
The volume of unemployed entrants and reentrants is significant in relation to
total unemployment. For this reason, unemployed entrants and reentrants are
estimated as a separate component of unemployment by means of a special
estimation procedure.

LAUS Program Manual 7-19

Labor Market Area Unemployment

Method for Estimating Unemployed Entrants and Reentrants

The new entrant and reentrant estimate is obtained by relating monthly
national seasonal factors, adjusted to reflect the relative concentration of
youth in the estimating area, to the experienced unemployed and the
experienced labor force in the estimating area.
Each year at benchmark time, the Youth Population Ratio (YPR) is
calculated for each area using the cohort method of survival and survival
rates derived from the Department of Health and Human Services U.S.
Life Tables. These survival rates are applied to individual age categories
to estimate the number of survivors in a given year. After the number of
survivors is computed, estimates of the 16-19 age group population and
the 20+ population for the given year are made and the YPR is derived.
The YPR is the ratio between the 16- to 19-year-old population and the
population 20 years of age and over. It is used throughout the year.
Survived 16-19 years old

97,205

÷ Survived 20 years and over

1,458,565

Youth population ratio (YPR)

6.7%

Regression Formulas for Annual A and B Factors. The relationship of
the Youth Population Ratio both to the ratio of new worker unemployed to
the experienced labor force and to the ratio of new worker unemployed to
the experienced unemployed has been extensively researched. To reduce
the effect of irregular changes in the data, the two new-worker ratios were
smoothed by the application of a five-year moving average. Two
functional relationships were then set up, with logarithmic equations
selected as the most suitable for explaining the functional relationships.
The analysis indicated a continued reasonable correlation between the
YPR and the ratio of new-worker unemployed to the experienced
unemployed (0.85 correlation coefficient) and the ratio of new-worker
unemployed to the experienced labor force, excluding new-worker
unemployed (0.70). The following equations were developed:

Y1 = -.019885 + .011151lnX
Y2 = -.3987 + .2271 lnX

7-20 LAUS Program Manual

Labor Market Area Unemployment

where:
X = Youth population ratio
Y1 = Ratio of new entrant unemployed to the civilian labor force less new entrant
unemployed (A factor)
Y2 = Ratio of new entrant unemployed to experienced unemployed (B factor)

Tables 7-4 and 7-5 list the annual A and B factors for YPRs of 6.5 through 17.5 as
computed from the above equations. For YPRs below 6.5, the annual A and B
factors for a YPR of 6.5 should be used. For any YPR above 17.5, the annual A
and B factors should be computed using the above equations.
High Unemployment “B” Factor. To reflect the lower labor force participation
rates and the discouraged worker effect in high unemployment areas, a separate
formula has been developed for the estimation of a high unemployment B factor.
The formula using the natural log of the YPR as in the above section with the
coefficients at an appropriately lower level. Using the same notation as above, the
equation is as follows:
Y2 = -.2868 + .1634 lnX
When the average unemployment rate in the previous 12 months reaches 6.5
percent or more in an area, the “high” unemployment B factor is used in the
following month. When the twelve month moving average reaches 6 percent, use
of the “high” B factor is terminated.

Y2 = -.2868 + .1634 lnX

Estimation of Entrants. The Annual A and B factors for an area are selected by
using the area’s YPR. “A” factor unemployment is the estimate of that portion of
new and reentrant unemployment which is related to the level of the experienced
labor force. The month’s A factor is the product of the area’s Annual A factor and
the monthly seasonal A′ factor developed from national CPS data.
-0.019885+(0.011151 × lnYPR)

0.001325

× Seasonal A′

1.033000

A factor

0.001369

LAUS Program Manual 7-21

Labor Market Area Unemployment

“B” factor unemployed is the estimate of that portion of new and
reentrant unemployment which is related to the level of experienced
unemployed. The month’s B factor is the product of the area’s Annual B
factor and the monthly seasonal B′ factor.
-0.3987+(0.2271 × lnYPR)

0.033269

× Seasonal B′

1.101100

B factor

0.036632

Seasonal monthly A‘ and B‘ factors are developed annually from CPS
data and provided to the States. (See tables 7-4 and 7-5 for factors from
1990 onward.) These seasonal monthly factors are multiplied by the
respective Annual A and B factors to yield the monthly A and B factors.
The month’s A factor is multiplied by the number of employed plus
unemployed excluding entrants. The resulting estimate is the number of
entrants related to the experienced labor force.
Employed

951,965

+ Unemployed, excluding entrant

35,630

Experienced labor force

987,595

× A factor

0.001369

A factor unemployed

1,352

The month’s B factor is multiplied by the number of unemployed
excluding entrants to produce the estimate of entrants related to the
experienced unemployed.
Unemployed, excluding entrants

35,630

× B factor

0.036632

B factor unemployed

1,305

States must determine, on a monthly basis, whether areas should use the
high or low unemployment B factor. When the average unemployment
rate for the preceding 12 months is 6.5 percent or greater, the high B
factor must be used. Otherwise, the Annual A and B factors appropriate
for the YPR are used for the entire year.

7-22 LAUS Program Manual

Table 7-1 STRATA STEP 3 RATIOS, BY MONTH

1990

STRATUM
1
2
3

JAN
1.03243
0.99663
0.99095

FEB
0.96902
0.99813
0.96557

MAR
0.98032
1.00937
0.99596

APR
1.01964
0.99068
1.00402

MAY
1.01098
0.99968
0.98982

JUN
1.03678
1.00725
0.98795

1
2
3

JAN
1.02376
1.00602
0.97780

FEB
1.00152
0.99897
0.99974

MAR
1.02225
1.01705
1.01593

APR
1.07187
1.03655
1.08472

MAY
1.01362
1.00949
0.98483

JUN
1.06922
1.05968
0.98871

1
2
3

JAN
1.04350
0.97031
0.94585

FEB
1.06728
1.00297
0.90717

MAR
1.05761
1.00587
0.94418

APR
1.09677
0.96574
0.94501

MAY
1.08101
0.98827
0.92798

JUN
1.13366
1.00940
0.94208

1
2
3

JAN
1.12069
0.98321
0.92885

FEB
1.09586
1.01558
0.94054

MAR
1.07738
1.00789
0.96657

APR
1.12774
1.02530
0.94147

MAY
1.16527
1.02025
1.00648

JUN
1.15812
1.03469
0.99355

1
2
3

JAN
1.12000
0.97300
0.81600

FEB
1.13700
1.01000
0.89000

MAR
1.13800
0.99800
0.86600

APR
1.13400
0.98400
0.84100

MAY
1.10500
0.98700
0.92900

JUN
1.06800
1.00800
0.89900

1
2
3

JAN
1.10500
0.94800
0.83000

FEB
1.08000
0.95000
0.84100

MAR
1.06800
0.94000
0.84400

APR
1.07400
0.91500
0.84900

MAY
1.05100
0.91800
0.85900

JUN
1.02500
0.92900
0.85000

JUL
1.05290
0.99964
1.03005

AUG
1.05038
0.99590
1.02865

SEP
1.05951
0.99133
0.98239

OCT
1.08174
1.01251
1.01556

NOV
1.01929
1.00550
0.99169

DEC
1.01201
1.01533
1.01199

AUG
1.13793
1.03557
0.99690

SEP
1.08895
1.01866
0.98801

OCT
1.14263
1.03430
0.99935

NOV
1.10022
1.01405
0.98286

DEC
1.07805
0.98472
0.97105

AUG
1.10997
1.01227
1.01393

SEP
1.10200
0.99513
0.98947

OCT
1.07218
0.99699
0.98178

NOV
1.08245
1.00930
0.96194

DEC
1.06903
1.00506
0.91902

AUG
1.19589
1.05684
0.99626

SEP
1.19726
0.98949
0.99669

OCT
1.22469
1.00757
0.97670

NOV
1.16273
1.00795
0.95740

DEC
1.18538
0.99129
0.92547

AUG
1.11900
0.99500
0.90100

SEP
1.07000
0.95700
0.90500

OCT
1.10300
0.94800
0.85000

NOV
1.10200
0.94400
0.88600

DEC
1.11500
0.95000
0.84200

AUG
1.06200
0.95000
0.85000

SEP
1.06600
0.93100
0.86600

OCT
1.09100
0.92800
0.83200

NOV
1.08600
0.90400
0.81800

DEC
1.10000
0.91300
0.78900

1991
JUL
1.10538
1.06765
0.97965

1992
JUL
1.15729
1.01131
1.04440

1993
JUL
1.21064
1.04254
1.00028

1994

1995
JUL
1.10600
0.96300
0.85900

Revised 02/24/03

Monthly Step 3 Ratios

LAUS Program Manual 7-23

JUL
1.09000
0.99900
0.93000

Table 7-1 STRATA STEP 3 RATIOS, BY MONTH

1
2
3

JAN
1.07500
0.90900
0.77100

FEB
1.08300
0.91800
0.79200

MAR
1.08800
0.91700
0.82500

APR
1.08500
0.89700
0.82400

MAY
1.03200
0.89400
0.85300

JUN
1.02300
0.91100
0.85900

1
2
3

JAN
1.09100
0.96000
0.81300

FEB
1.03600
0.92200
0.80600

MAR
1.05900
0.92500
0.82800

APR
1.04600
0.90900
0.82400

MAY
1.02600
0.90400
0.82800

JUN
1.00500
0.92600
0.80700

1
2
3

JAN
0.99300
0.89900
0.77900

FEB
0.96600
0.88200
0.76700

MAR
1.00500
0.88400
0.74900

APR
1.02000
0.89300
0.77900

MAY
1.02200
0.88800
0.76400

JUN
1.01400
0.87300
0.78200

1
2
3

JAN
0.94600
0.85500
0.74000

FEB
0.91800
0.82600
0.72700

MAR
0.93900
0.84500
0.72600

APR
0.92600
0.85300
0.74300

MAY
0.95300
0.84800
0.71500

JUN
0.97100
0.84400
0.77500

1
2
3

JAN
0.91800
0.82400
0.69800

FEB
0.93600
0.83300
0.69900

MAR
0.88100
0.84600
0.72500

APR
0.90000
0.84700
0.74100

MAY
0.90700
0.82200
0.72800

JUN
0.91400
0.82700
0.71600

1
2
3

JAN
0.88400
0.82200
0.67200

FEB
0.85700
0.79300
0.68000

MAR
0.86900
0.82000
0.70800

APR
0.84400
0.80200
0.71700

MAY
0.86400
0.77300
0.68600

JUN
0.89100
0.80500
0.69200

1
2
3

JAN
0.84000
0.76000
0.64500

FEB
0.83600
0.76200
0.63500

MAR
0.81700
0.77000
0.66300

APR
0.85900
0.76200
0.71700

MAY
0.87500
0.74800
0.70200

JUN
0.85200
0.78200
0.68300

JUL
1.01400
0.94600
0.87500

AUG
1.00000
0.92100
0.90800

SEP
0.99400
0.89800
0.85700

OCT
1.01400
0.92800
0.85300

NOV
1.03600
0.91900
0.80600

DEC
1.04500
0.93100
0.79700

AUG
1.03100
0.90900
0.80000

SEP
0.98700
0.88600
0.78700

OCT
0.99000
0.88100
0.79200

NOV
0.98300
0.89700
0.77100

DEC
0.97200
0.88700
0.77200

AUG
1.06000
0.89400
0.77800

SEP
1.01200
0.87600
0.75800

OCT
0.99500
0.87900
0.76800

NOV
0.96800
0.86600
0.76100

DEC
0.97800
0.83300
0.74500

AUG
0.95900
0.86500
0.76600

SEP
0.97600
0.82400
0.77900

OCT
0.92300
0.85000
0.76400

NOV
0.90200
0.82400
0.74400

DEC
0.90100
0.81100
0.72400

AUG
0.91500
0.79700
0.69100

SEP
0.09170
0.81300
0.71500

OCT
0.90600
0.82200
0.71300

NOV
0.91400
0.76200
0.68500

DEC
0.89500
0.78800
0.68400

AUG
0.90700
0.77100
0.67100

SEP
0.91000
0.77500
0.67900

OCT
0.89700
0.77800
0.68200

NOV
0.88900
0.75500
0.65700

DEC
0.85800
0.76500
0.65000

AUG
0.89200
0.79300
0.66300

SEP
0.92100
0.79000
0.69600

OCT
0.89700
0.80200
0.73500

NOV
0.94600
0.81600
0.72300

DEC
0.86100
0.82100
0.69100

1997
JUL
0.99500
0.92700
0.80800

1998
JUL
1.06000
0.89400
0.77800

1999
JUL
0.95300
0.86800
0.76100

2000
JUL
0.90200
0.81700
0.71800

2001
JUL
0.91100
0.77800
0.71100

2002
JUL
0.91000
0.77100
0.70200

Revised 02/24/03

Monthly Step 3 Ratios

7-24 LAUS Program Manual

1996

STRATUM

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.65195
0.85911
0.87058
0.80085
0.71514
0.77051
0.94126
0.82954
0.58676
0.60750
1.07437
1.11434
1.12476
0.65353

0.78521
0.85180
0.93949
0.78093
0.79601
0.82250
0.76903
0.78209
0.61523
0.75846
1.04611
1.00424
0.96187
INA

0.77559
0.89473
0.96322
0.69944
0.76796
1.00574
0.92081
0.91548
0.68859
0.86902
1.21242
0.81579
0.61377
INA

1.00000
0.98821
0.91227
0.84582
1.08756
1.10251
1.04809
1.09422
0.99842
1.17538
1.46105
0.94436
0.79901
INA

1.27118
1.16658
0.98312
1.02685
1.27144
1.54339
1.40590
1.21866
1.04645
1.19663
1.31458
0.90640
0.82415
INA

1.24191
1.43340
1.13356
1.19425
1.43437
1.51532
1.38475
1.45373
1.03967
1.28162
1.56544
1.11030
1.10530
INA

1.19274
1.31302
1.24286
1.35311
1.31302
1.33307
1.36314
1.37316
1.26291
1.32304
1.51348
1.39321
1.18272
INA

1.10560
1.36424
1.23220
1.25498
1.34698
1.24503
1.17574
1.27234
1.02851
1.24221
1.53493
1.50700
1.23128
INA

1.00163
1.06980
1.09153
1.19107
1.11432
1.15936
1.15460
1.26707
1.03975
1.20071
1.45777
1.53485
1.20522
INA

0.91260
0.99201
1.05646
1.11919
1.08678
1.16310
1.07267
0.99743
0.92965
1.20433
1.34638
1.41790
1.19420
INA

0.88214
1.16689
1.05963
1.21713
1.14562
1.02976
0.88180
0.93016
0.73003
1.04992
1.19625
1.38838
1.15947
INA

0.92114
1.01118
0.85042
1.01007
0.96627
1.13445
0.86156
0.68778
0.69274
1.01066
1.24136
1.26304
1.03306
INA

NORTHEAST II

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.88347
0.70319
0.67578
0.62168
0.58836
0.88857
0.95246
0.96734
0.63544
0.78597
0.91971
0.75871
0.97048
0.56344

0.95733
0.69023
0.64413
0.69514
0.61363
0.89233
0.75812
0.65256
0.77041
0.83847
0.92128
0.96424
0.94661
INA

0.98073
0.78055
0.74393
0.80331
0.81765
0.98506
0.68596
0.87213
0.84838
0.73445
1.09331
0.96331
0.89518
INA

1.00000
0.93779
0.88384
1.02846
0.92626
0.97723
0.65267
0.91434
1.08848
0.95809
1.23126
1.11893
0.95171
INA

1.18535
0.87245
1.06049
0.97784
0.90995
0.88361
0.82358
1.16143
1.22747
0.94583
1.33278
1.22147
0.99307
INA

1.12983
0.91123
1.12422
1.11150
1.08236
1.07657
0.95514
0.98795
1.19393
1.04837
1.29440
1.33481
1.11556
INA

1.18676
1.06808
1.10517
1.07550
1.02358
1.19418
1.15709
1.30544
1.24610
1.13484
1.28318
1.27577
1.17193
INA

1.12996
1.03942
0.97358
0.90283
0.97929
1.33808
1.36726
1.36722
1.18685
1.20835
1.41830
1.20633
1.13442
INA

1.10962
1.01539
0.89868
0.80243
1.06829
1.12184
1.46421
1.43448
1.18527
1.03191
1.33533
1.08387
1.17650
INA

1.07067
0.99611
0.81744
0.74417
1.02797
1.16202
1.36308
1.30986
1.03658
0.96473
1.27734
1.07706
1.28353
INA

0.87590
0.93055
0.65408
0.75400
1.02300
1.24099
1.27570
1.05733
1.01531
0.98607
1.17751
1.00870
1.14030
INA

0.83132
0.72343
0.57437
0.78110
1.04650
0.98076
1.19196
0.83544
0.87836
0.89332
1.17066
1.14270
1.13747
INA

Revised 2/24/03

Monthly Agricultural Factors

LAUS Program Manual 7-25

NORTHEAST I

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

APPALACHIAN I

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.75455
0.58234
0.67595
0.78453
0.45392
0.68046
0.57844
0.53548
0.48345
0.79510
0.50499
0.75647
0.68628
0.52616

0.75550
0.77820
0.81142
0.79915
0.64462
0.70800
0.48750
0.62597
0.34818
0.70613
0.49046
0.90270
0.70483
INA

0.89166
0.91980
0.80741
0.72585
0.80838
0.97482
0.54360
0.67410
0.45252
0.92171
0.72451
1.01981
0.62512
INA

1.00000
1.01556
0.97661
0.85773
0.58055
0.93071
0.53158
0.69711
0.65688
1.00293
0.64593
0.95710
0.67093
INA

1.10703
1.26132
0.92497
0.93644
0.85218
0.95925
0.81954
0.89541
0.68912
0.82402
0.73964
0.94539
0.85155
INA

1.13145
1.09567
1.00850
0.86202
0.71534
1.02195
0.70587
0.96234
0.87923
1.01778
0.86920
0.99997
1.01014
INA

0.93969
1.15401
1.01113
0.85177
0.87925
1.03311
0.91222
0.93420
0.87375
1.01663
0.90672
0.97816
0.94519
INA

0.82687
1.09275
0.94938
1.02312
0.92818
1.03138
1.04051
0.93458
0.82838
1.05023
1.00656
1.31424
0.83270
INA

0.91721
1.02108
0.95420
0.89324
0.94625
0.91107
0.88221
0.83282
0.86400
1.21952
0.91849
1.29804
0.89286
INA

0.82443
0.80057
0.88242
0.74438
0.86047
0.83173
0.85072
0.86615
0.90120
1.11078
0.89188
1.18370
0.89821
INA

0.90019
0.81499
0.73263
0.64825
0.72084
0.86845
0.79279
0.72773
0.81900
0.76858
0.82731
1.01775
0.92792
INA

0.70004
0.71070
0.63551
0.58541
0.67019
0.67276
0.70387
0.63884
0.80341
0.50796
0.80282
0.81455
0.95502
INA

APPALACHIAN II

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.90779
0.84489
0.86714
0.84975
0.72469
1.02396
1.13026
1.30702
0.71450
0.68054
0.72693
0.83142
1.26180
1.02919

1.12613
0.87377
0.88100
0.82422
0.76947
1.03371
1.09542
0.76825
0.92119
0.82687
0.78538
0.90519
1.45404
INA

1.14312
0.76452
0.85946
0.84557
0.78754
1.05942
1.18333
1.11920
1.01714
1.17288
1.01476
1.19286
1.54174
INA

1.00000
0.69850
0.78270
0.75339
0.89399
1.04209
1.07274
1.41693
1.18372
1.19412
0.91377
1.21250
1.62624
INA

1.07084
1.08663
0.97507
1.08466
0.98712
1.06919
1.22582
1.16398
1.30111
1.62500
1.06530
1.26404
1.23096
INA

1.37003
1.19307
1.25238
1.15195
0.91765
1.12886
1.15602
1.23044
1.34011
1.51825
1.17080
1.21232
0.95807
INA

1.08829
1.07797
1.23787
1.12955
0.94903
1.16050
1.20692
1.27913
1.15018
1.22755
1.19660
1.18113
0.96966
INA

1.35496
1.24579
1.49185
1.00095
1.03654
0.91648
1.08501
1.02698
1.38406
1.18280
1.17653
1.02455
0.96535
INA

1.25482
1.20270
1.21897
1.18106
1.14497
1.00504
1.08832
0.79583
1.23331
0.98454
1.20353
1.12437
0.83919
INA

1.33240
1.09290
1.04492
0.93378
1.01695
1.06869
1.21862
0.84294
0.82917
0.82060
1.12008
1.20750
1.17300
INA

1.29697
1.30374
1.17811
0.90892
0.96365
1.30043
1.32256
0.86927
0.95871
0.57806
1.01502
1.24713
0.97970
INA

1.15244
1.09983
1.09375
0.73352
1.21959
1.16717
1.11303
0.76321
0.83647
0.81267
0.88878
1.46775
1.05113
INA

Revised 2/24/03

Monthly Agricultural Factors

7-26 LAUS Program Manual

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.99523
0.84476
0.75529
0.86738
0.81877
0.68584
0.71469
0.94686
0.64707
0.76902
1.28249
1.06085
0.52365
0.47277

0.94439
0.79521
0.78448
0.76054
0.51754
0.73625
0.90349
1.00804
0.75948
0.75093
1.23034
0.75819
0.44110
INA

1.07607
1.05011
0.95501
0.86318
0.69143
0.76832
0.96726
1.13673
0.60828
0.72269
1.10881
0.99355
0.44689
INA

1.00000
1.10164
1.05296
0.75150
0.70147
0.91528
0.95671
1.12717
0.44902
0.72953
1.10314
1.01053
0.57654
INA

1.14080
1.07848
1.02190
0.78510
0.94300
0.89282
1.05238
1.03487
0.50771
0.74161
1.06890
1.24806
0.63291
INA

1.23112
1.13367
1.16398
1.12185
0.98063
1.00512
1.09115
1.06339
0.72648
0.96599
1.12355
0.90506
1.00040
INA

1.17455
1.21916
1.15225
1.14482
0.98871
1.03331
0.98871
1.12995
0.95897
1.01101
1.12995
1.01844
1.08535
INA

1.00170
0.84565
1.18676
0.99324
0.95682
0.82055
0.79849
1.00536
0.75883
0.83682
1.10271
0.95920
0.83250
INA

1.06771
0.91432
0.91108
1.14088
0.70008
0.60273
0.72953
0.88652
0.82524
0.88015
1.25381
0.92562
0.75372
INA

1.01115
0.91541
0.88133
0.87442
0.66270
0.66991
0.76899
0.71620
0.76871
0.92150
1.34426
1.04990
0.76660
INA

0.82344
0.76175
0.93463
0.78523
0.76302
0.57749
0.69495
0.68301
0.66527
1.07116
1.20356
0.65218
0.66302
INA

0.84271
0.72434
0.82309
0.91388
0.77296
0.74541
0.93912
0.78586
0.79478
1.20808
1.03647
0.72449
0.65218
INA

FLORIDA

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

1.24595
1.42849
1.13372
1.35880
1.52417
1.49139
1.96518
1.72701
1.39399
1.55024
1.83768
1.97551
1.46832
0.47537

0.89573
1.31441
1.04311
1.18200
1.54776
1.62219
2.13012
1.61884
1.31867
1.43725
1.56881
1.75748
1.44618
INA

1.16354
1.27557
1.15935
1.19023
1.40701
1.80181
2.08248
1.62087
1.14081
1.61765
1.59567
1.73236
1.39729
INA

1.00000
1.18548
1.47371
1.04769
1.29851
1.73763
2.15551
1.84173
1.33383
1.27559
1.60430
1.74425
1.37041
INA

1.15502
1.47866
1.03823
1.17496
1.15637
1.63677
1.64702
1.66144
1.29538
1.18425
1.39825
1.78906
1.40484
INA

1.10062
1.39747
1.16327
1.08166
1.21426
1.54544
1.40196
1.25286
1.26530
1.30992
1.57479
1.61058
1.18026
INA

1.10520
1.23917
1.12195
1.28940
1.17218
1.32289
1.37313
1.30615
1.33964
1.37313
1.35639
1.45686
1.30615
INA

0.97257
1.30040
1.09922
1.09969
1.21129
1.30375
1.14398
1.23089
1.46650
1.45752
1.37414
1.82352
1.31960
INA

1.07617
1.43864
1.32436
1.15896
1.20135
1.61154
1.39597
1.22291
1.71673
1.50764
1.32475
1.84518
1.62448
INA

0.98289
1.36171
1.22040
1.12089
1.37498
1.85474
1.24417
1.06111
1.92325
1.47553
1.48689
1.60864
1.51018
INA

1.22607
1.53531
1.15072
1.34663
1.65232
1.68850
1.60035
1.00491
1.48218
1.60875
1.57443
1.58566
1.38347
INA

1.37093
1.38247
1.51527
1.44356
1.52604
1.73881
1.70416
1.25839
1.22815
1.74692
1.62370
1.16280
1.24048
INA

Revised 2/24/03

Monthly Agricultural Factors

LAUS Program Manual 7-27

SOUTHEAST

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

LAKE

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

1.19156
0.97526
0.74215
0.88231
0.94178
0.74756
0.76390
0.69132
0.70532
0.64242
0.59324
0.45744
0.53673
0.43641

1.09421
0.87076
0.70925
0.83695
0.93716
0.74910
0.84259
0.71506
0.78630
0.64658
0.71026
0.38267
0.57763
INA

1.12142
0.83387
0.62788
0.87401
0.89680
0.84610
0.86057
0.62803
0.70520
0.54269
0.77552
0.45780
0.60581
INA

1.00000
0.87898
0.81501
0.95182
0.84217
0.99139
0.88521
0.73356
0.78276
0.62570
0.83757
0.72037
0.59508
INA

1.10103
1.00741
0.97285
1.01506
1.00976
0.96289
0.98450
0.81604
0.92516
0.73887
0.90430
0.76147
0.62494
INA

1.08062
1.09576
1.06486
1.10344
1.09227
1.04655
0.91270
0.80197
0.86724
0.72965
0.86387
0.81886
0.62260
INA

1.16534
1.14288
0.98282
1.08391
1.05021
1.02775
0.96597
0.88734
0.97720
0.94070
0.90981
0.83680
0.68236
INA

1.25954
1.05855
0.97659
1.00377
1.22234
0.98699
0.97768
0.98907
0.83088
0.90528
0.79402
0.88663
0.64495
INA

1.19026
0.94713
0.86295
1.00631
1.20554
0.91333
0.94581
0.96649
0.74773
0.75246
0.79480
0.79099
0.75575
INA

1.19635
1.03234
0.92479
1.00704
1.13857
0.87775
0.90132
0.97645
0.81258
0.77731
0.80507
0.72931
0.83107
INA

1.06413
0.86683
0.88823
0.96561
1.11286
0.86200
0.85400
0.93587
0.70526
0.66448
0.67048
0.58323
0.69849
INA

1.04160
0.78851
0.79450
0.90508
0.78782
0.76953
0.73662
0.90811
0.70606
0.56858
0.52514
0.56760
0.61192
INA

CORN BELT I

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.69185
0.64476
0.58084
0.65220
0.80575
0.90080
0.69513
0.67544
0.63119
0.85625
0.78161
0.58651
0.48397
0.37446

0.60722
0.63436
0.58581
0.70547
0.81272
0.92445
0.64487
0.60511
0.54915
0.85242
0.72631
0.68716
0.40452
INA

0.74982
0.73071
0.57201
0.80146
0.81208
0.91155
0.71167
0.72974
0.56087
0.92971
0.71447
0.81694
0.52278
INA

1.00000
0.84283
0.63763
0.89603
0.96435
1.03270
0.82170
0.77096
0.78293
0.95878
0.79760
0.78089
0.63979
INA

1.11327
0.98744
0.86140
1.01215
0.98537
1.03411
0.88247
0.81534
0.80581
1.12361
0.88880
0.83052
0.66178
INA

1.23332
1.06787
0.97525
1.06091
1.11929
1.21129
1.05827
1.11972
1.01308
1.19912
0.88438
1.02158
0.76094
INA

1.18499
1.04895
0.99525
1.10981
1.10623
1.09549
1.06685
1.05611
0.92365
1.09549
0.99883
1.06685
0.83415
INA

1.09280
0.89931
0.83121
0.98967
0.90212
1.09901
0.99174
0.85883
0.91047
1.14739
1.06735
1.00121
0.70517
INA

0.84060
0.88702
0.74169
0.89139
0.92852
1.09188
0.99905
0.83682
0.85586
0.97731
0.86934
0.92611
0.67371
INA

0.80733
0.81658
0.77796
0.98027
1.00025
1.07545
0.92962
0.78870
0.94929
0.81598
0.69472
0.82403
0.69698
INA

0.75437
0.67117
0.73946
0.96154
1.04951
1.07562
0.82307
0.79868
0.93796
0.71040
0.64966
0.68684
0.59610
INA

0.64815
0.58504
0.69623
0.76394
1.04411
0.96162
0.76589
0.74604
0.90335
0.55013
0.60336
0.61335
0.62132
INA

Revised 2/24/03

Monthly Agricultural Factors

7-28 LAUS Program Manual

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.72869
0.87549
0.87616
0.63823
0.77629
1.00820
0.82354
0.91360
0.98853
0.96092
0.79758
0.91778
1.06515
0.66360

0.83639
0.90518
0.92505
0.59592
0.80987
0.92602
0.83446
0.85499
0.71972
0.92021
0.79197
1.16303
1.02532
INA

0.93717
0.76884
0.88199
0.68386
0.87926
1.05392
1.06496
0.88084
0.70163
1.05115
0.74095
1.05184
0.89004
INA

1.00000
0.83809
0.85781
0.65600
0.93585
0.93004
1.04955
0.97149
0.78098
1.09457
0.85392
1.20633
0.96902
INA

1.11076
0.90804
0.87033
0.70465
1.01015
1.01490
1.14290
1.10962
0.85748
1.08796
1.07561
1.22838
0.81359
INA

1.16373
0.92605
0.95669
0.81626
1.01162
1.05420
1.11931
1.09326
0.92832
1.18047
1.00169
0.96690
0.84550
INA

1.05444
1.06679
1.04208
0.98442
0.96794
1.03796
1.07503
1.05032
0.93499
1.09563
1.02149
1.08739
1.12446
INA

0.99813
0.93708
1.03796
0.90302
0.95512
0.98905
1.08159
1.05289
0.88120
1.04606
0.96606
1.07933
1.11578
INA

0.82384
0.94064
0.94335
0.85511
0.92399
0.83417
1.02667
1.06238
0.93659
1.06160
0.89190
0.96550
1.11385
INA

1.01003
0.94401
0.95689
0.94721
0.97720
0.88070
1.10014
1.16426
1.04166
1.05436
0.81910
1.15225
1.06867
INA

1.01591
0.95511
0.79153
0.77216
1.00994
0.77043
0.95901
1.09872
0.91575
1.03870
0.89601
1.08027
0.88768
INA

0.91785
0.89722
0.72036
0.68816
1.04935
0.76238
0.85739
1.08669
0.98788
0.95149
0.70222
0.93476
0.77424
INA

DELTA

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.97475
0.85715
1.14763
1.33862
1.30624
1.36860
1.01522
0.86046
1.00397
1.04456
1.14248
0.76297
1.10624
1.40554

0.86557
1.00396
1.28151
1.32859
1.43233
1.32559
0.96310
1.05385
1.00299
0.80327
0.93216
0.86667
0.96027
INA

0.91638
1.04760
1.35229
1.52028
1.81058
1.28279
1.04927
1.06482
0.97000
0.91481
0.98272
1.10633
1.11409
INA

1.00000
1.09360
1.38792
1.51806
1.84461
1.26274
1.43527
1.13875
0.89486
1.05550
1.24008
1.08011
1.12353
INA

1.26455
1.25691
1.39280
1.61604
1.69172
1.26151
1.14181
1.15658
0.95808
1.11561
1.31380
1.22358
1.35796
INA

1.26078
1.38048
1.32343
1.65993
1.45073
1.56202
1.21802
1.31919
1.07296
1.26968
1.23047
1.30795
1.39977
INA

1.23479
1.36006
1.35111
1.50322
1.29742
1.37795
1.43164
1.35111
1.26163
1.39585
1.51217
1.41374
1.41374
INA

1.29382
1.31078
1.21916
1.50134
1.20632
1.34643
1.03758
1.16318
1.12610
1.41034
1.47300
1.06909
1.49851
INA

1.25838
1.10844
1.26060
1.26990
1.20587
1.38237
1.27000
1.17356
1.05517
1.24882
1.52408
1.26302
1.47847
INA

1.22291
1.32897
1.37252
1.29013
1.42895
1.39245
1.11384
0.98400
1.05867
1.33021
1.53797
1.43749
1.80030
INA

1.10766
1.32001
1.34952
1.20355
1.44396
1.02100
0.87579
0.84541
1.08193
1.27642
1.35067
1.51388
1.60206
INA

0.88165
1.12287
1.39880
1.36719
1.36034
0.95049
0.78414
0.86509
1.09067
1.05925
1.20952
1.15104
1.49579
INA

Revised 2/24/03

Monthly Agricultural Factors

LAUS Program Manual 7-29

CORN BELT II

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

NORTHERN PLAINS

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.88914
0.91825
0.80666
0.84307
0.92697
0.97864
0.96047
0.84267
0.76442
0.93676
0.88843
0.85958
0.92812
0.83560

0.91152
0.89154
0.85092
0.91625
0.99577
1.03189
0.95336
0.77767
0.74372
0.92896
0.94662
0.85285
0.86825
INA

0.96767
0.93882
0.89097
0.92492
1.02225
1.04439
0.96655
0.73535
0.76541
0.90382
0.93002
0.90197
0.84776
INA

1.00000
1.01178
0.92116
1.01202
1.05631
1.07275
0.98218
0.86586
0.92465
0.92100
0.92343
0.95118
0.92407
INA

1.05198
1.07098
1.00529
1.06073
1.06210
1.10369
1.05726
0.98745
0.97964
1.08892
1.09806
1.00039
1.08756
INA

1.19957
1.13342
1.00512
1.06024
1.10280
1.25427
1.07292
0.99621
1.04914
1.13925
1.20576
1.06165
1.09924
INA

1.22183
1.14066
1.07658
1.01677
1.06804
1.16202
1.06376
1.08085
1.09367
1.23465
1.09367
1.06804
1.09794
INA

1.17310
1.16364
1.08448
0.97447
1.08375
1.18675
1.06043
0.99003
1.10252
1.22038
1.09158
0.97762
1.01303
INA

1.08912
1.06954
1.05110
0.89464
1.00683
1.09413
1.06343
0.98151
0.96188
1.09720
0.95740
0.95103
0.91024
INA

1.04092
1.00779
1.04835
0.87569
1.05362
1.13325
1.08262
0.94482
0.94501
1.03376
0.96401
1.00008
0.99753
INA

0.94517
0.89044
0.96841
0.84944
1.02140
1.05829
0.99304
0.83283
0.96067
0.96320
0.92605
0.94866
0.95447
INA

0.90536
0.85346
0.87366
0.80069
1.01191
1.04967
0.91637
0.83765
0.94194
0.93375
0.88032
0.85492
0.84218
INA

SOUTHERN PLAINS

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.75824
1.07000
0.94674
0.79426
0.87222
1.10725
1.08565
1.10099
1.26749
1.10626
1.05557
1.14536
1.10654
0.84353

0.90452
1.06522
0.92278
0.91901
0.93114
1.02694
1.04303
1.10692
1.21361
1.17550
1.07182
1.07336
1.01971
INA

0.96091
0.97034
1.07374
0.94424
0.92417
1.09408
1.11337
1.24017
1.18524
1.14042
1.14578
1.19741
1.20342
INA

1.00000
0.96237
1.11510
1.03737
1.26261
1.14590
1.21460
1.22176
1.19769
1.31435
1.08283
1.21221
1.20533
INA

1.06516
0.99196
1.20617
1.16207
1.26873
1.09649
1.22596
1.16580
1.41529
1.41089
1.18819
1.30019
1.23726
INA

1.20400
1.13443
1.24460
1.30583
1.19635
1.12997
1.33924
1.19067
1.50663
1.47752
1.30797
1.35872
1.23168
INA

1.11346
1.05888
1.20079
1.22990
1.24809
1.25537
1.31359
1.29540
1.32815
1.47006
1.49553
1.58286
1.37909
INA

1.05965
1.07668
0.95854
1.19960
1.27185
1.12343
1.17744
1.21466
1.37223
1.16181
1.36974
1.45005
1.26253
INA

0.96903
0.96777
0.91340
1.09638
1.25051
1.08654
1.10626
1.15071
1.17357
1.05077
1.35507
1.57295
1.32932
INA

1.09873
0.96059
0.86674
0.98227
1.24810
1.16486
1.10534
1.15702
1.16889
1.18492
1.26622
1.44155
1.22105
INA

1.08105
1.10632
0.91781
1.06466
1.25053
1.12252
1.04307
1.24209
1.03095
1.17555
1.19577
1.24235
1.11910
INA

1.20559
1.03744
0.84457
0.70632
1.16208
1.06373
1.12743
1.27467
1.03739
1.12062
1.29682
1.25498
1.00858
INA

Revised 2/24/03

Monthly Agricultural Factors

7-30 LAUS Program Manual

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.88207
0.80573
0.90140
0.87524
1.01383
0.88572
1.07261
0.86996
0.94199
1.41743
1.53369
1.03501
0.97184
0.94661

0.78510
0.77827
0.96708
0.81876
0.97097
0.98405
1.03831
0.80601
0.96289
1.46073
1.36433
0.93326
0.88381
INA

0.91381
0.93792
1.13456
0.89123
1.10133
1.01164
1.19664
0.95072
0.98004
1.39148
1.44324
0.96947
0.78731
INA

1.00000
1.07061
1.19507
0.97941
1.20359
1.12872
1.25221
0.98674
1.11725
1.47443
1.60450
1.16163
0.97675
INA

1.15849
1.08900
1.14548
1.06584
1.25764
1.10422
1.30504
1.14081
1.38237
1.49983
1.61953
1.16822
1.20455
INA

1.17862
1.27484
1.22467
1.16883
1.27070
1.14823
1.31645
1.26520
1.40899
1.54571
1.60560
1.29733
1.43023
INA

1.27427
1.32850
1.30139
1.20649
1.26072
1.22005
1.38272
1.34205
1.34205
1.53184
1.57251
1.50473
1.32850
INA

1.20441
1.22170
1.22380
1.11042
1.16776
1.27806
1.32892
1.41169
1.43393
1.52438
1.56869
1.46396
1.41440
INA

0.99058
1.10555
1.20992
1.11357
0.98778
1.15703
1.28800
1.46175
1.40164
1.51744
1.69522
1.44007
1.29312
INA

1.05434
1.16030
1.06078
1.00584
1.05203
1.15945
1.27406
1.31612
1.38432
1.74054
1.70292
1.39579
1.21146
INA

0.91437
0.88375
1.00451
0.87218
0.81349
0.99636
1.10429
1.19177
1.58696
1.79876
1.22419
1.19957
1.15313
INA

0.85826
0.87633
0.99513
0.88223
0.80981
1.01729
1.00809
0.99228
1.36619
1.50384
1.17138
1.07208
0.99923
INA

MOUNTAIN II

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.60665
0.44622
0.54234
0.47495
0.33040
0.58583
0.86802
0.79033
1.03021
0.88341
0.65523
0.68953
0.87890
0.59312

0.71510
0.58948
0.50574
0.46795
0.31741
0.65466
0.82445
0.82038
0.94750
0.82743
0.64125
0.74567
0.71044
INA

0.82408
0.61607
0.54854
0.50601
0.24048
0.74398
0.80486
0.87133
1.06373
0.96778
0.85459
0.78202
0.71859
INA

1.00000
0.85338
0.65407
0.50384
0.34564
0.70127
0.87600
0.92020
1.00993
1.27156
1.21097
0.88963
0.87730
INA

1.03365
0.88434
0.71921
0.63445
0.52902
0.81619
0.81596
0.93375
1.02606
1.24258
1.03125
0.97105
0.98945
INA

1.08898
0.86105
0.85746
0.76511
0.57059
0.84037
0.87831
0.96544
1.04257
1.11185
0.84148
0.82111
0.88290
INA

0.91817
0.89302
0.84271
0.79240
0.83013
0.91817
0.83013
0.89302
0.99364
1.08168
0.96848
1.04395
0.83013
INA

0.81904
0.73257
0.91875
0.84106
0.85507
0.77447
0.83291
0.90341
1.24513
0.80457
0.76118
0.81893
0.92593
INA

0.73221
0.71075
0.74587
0.87081
0.87222
0.70757
0.84016
1.01068
0.85201
0.67809
0.88879
0.90363
0.80005
INA

0.57639
0.60970
0.51395
0.72112
0.76628
0.69280
0.92328
1.03865
0.91339
0.73096
0.98770
0.88392
0.86528
INA

0.46147
0.59817
0.52007
0.59777
0.74056
0.69976
0.78012
0.83666
0.89172
0.87059
0.77819
0.94537
0.73484
INA

0.53875
0.41661
0.59428
0.54282
0.62799
0.71383
0.68917
0.92506
0.75112
0.61567
0.70237
0.76983
0.76346
INA

Revised 2/24/03

Monthly Agricultural Factors

LAUS Program Manual 7-31

MOUNTAIN I

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

MOUNTAIN III

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.90759
0.81138
0.97165
0.66502
0.81514
0.93958
1.08284
1.24528
1.11135
1.00369
1.58759
1.71386
0.82090
0.70517

0.94864
0.81088
0.67895
0.65504
1.04143
1.52109
1.71681
1.21798
0.95561
0.93634
1.38808
1.06564
0.99625
INA

0.87337
0.75192
0.68989
0.59288
1.33470
1.40467
1.08284
1.06801
0.91957
1.22890
1.31296
0.97567
1.00276
INA

1.00000
0.75046
0.54218
0.67627
1.28691
0.84564
0.95413
0.98593
0.77509
1.37109
1.29324
1.16969
0.98382
INA

0.71487
0.72828
0.62642
1.13086
1.52215
1.07661
0.98988
1.44353
0.89926
1.11278
1.33826
1.35977
1.11394
INA

0.87763
1.01118
0.95670
0.93954
1.37091
1.13597
1.22892
1.22156
1.18793
1.08607
1.44442
1.30062
1.02110
INA

0.95338
1.04165
1.02400
0.95338
1.05931
1.07696
1.09462
1.18289
1.04165
1.02400
1.09462
1.07696
0.95338
INA

0.93004
0.98412
0.92393
0.76413
0.98951
1.34741
1.32613
1.14423
1.24227
1.12164
0.96852
0.76174
1.27420
INA

0.80673
1.07453
0.81794
0.62175
0.79405
0.96078
1.40064
1.09948
1.23193
1.12649
0.74457
0.81051
1.27751
INA

0.71913
1.11858
0.93341
0.75981
0.86046
0.89539
1.36811
1.34561
1.12192
1.37708
1.04823
0.83443
1.75728
INA

0.76116
1.13394
0.83121
0.59589
0.93439
1.08257
1.40514
1.47178
1.11457
1.65172
1.66431
0.74281
1.57768
INA

0.84191
1.15639
0.72961
0.70089
1.07798
1.16412
1.17698
1.33414
0.79201
1.37383
1.71833
0.85608
1.36388
INA

PACIFIC

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

1.01758
0.74269
0.86284
1.17935
1.01644
0.87605
1.23041
1.26137
1.53313
1.33840
1.36819
1.07314
1.14511
0.85470

0.83483
0.81040
0.74235
1.27283
0.99275
0.77669
1.15131
1.28226
1.28942
1.34819
1.57468
1.57106
1.19279
INA

0.92917
0.77418
0.63880
1.26580
0.80434
0.90937
1.35094
1.63091
1.62417
1.42421
1.83280
1.41090
1.35280
INA

1.00000
1.14344
0.83247
1.36159
0.95471
1.10894
1.16065
1.73864
1.49927
1.40444
1.69201
1.31910
1.37209
INA

1.41523
1.32680
0.97372
1.17835
0.84188
1.01973
1.39891
1.47250
1.15107
1.51302
1.57732
1.56634
1.59050
INA

1.61416
1.29856
1.10534
1.28634
1.01057
1.42598
1.55451
1.31700
1.21748
1.83701
1.63093
1.51664
1.37822
INA

1.60299
1.47475
1.22628
1.33849
1.44269
1.50681
1.65108
1.60299
1.35452
1.56291
1.41063
1.57093
1.58696
INA

1.60987
1.31774
1.14041
1.27074
1.48007
1.41005
1.34235
1.55166
1.12406
1.26385
1.84754
1.49673
1.49579
INA

1.52341
1.38685
1.20574
1.25917
1.28346
1.20605
1.40958
1.49050
1.76200
1.67223
1.82419
1.31262
1.26019
INA

1.16853
1.05103
1.28336
1.00385
1.06638
1.38769
1.35767
1.59515
1.89202
1.23337
1.27745
1.22822
1.49079
INA

0.88221
0.87094
1.49486
1.28502
0.96258
1.12746
0.99388
1.58483
1.46319
1.65954
1.62233
1.07557
1.10949
INA

0.79825
0.80937
1.24778
1.03334
0.94434
1.07071
0.99018
1.35086
1.33488
1.70358
1.49556
1.17442
1.14416
INA

Revised 2/24/03

Monthly Agricultural Factors

7-32 LAUS Program Manual

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

0.89084
0.86183
0.88892
1.14238
1.16107
0.95979
0.93620
0.96620
1.03673
1.12112
1.01563
1.18342
1.24208
0.62183

0.84343
0.88272
0.99411
1.00655
0.98826
1.05386
1.04030
0.98637
1.18680
1.19949
1.09894
1.11085
1.32184
INA

0.93347
0.89549
1.06747
1.08634
0.99685
0.99294
1.01944
0.99465
1.13936
1.18124
0.86876
1.08658
1.14956
INA

1.00000
1.02973
1.03421
1.17606
1.09442
1.10509
1.02008
1.10316
1.34403
1.37939
1.12342
1.24643
1.37476
INA

1.05525
1.06890
1.04025
1.15804
1.16531
1.11843
1.27202
1.26572
1.39732
1.51266
1.17926
1.39705
1.28302
INA

1.24096
1.32194
1.17036
1.22417
1.04933
1.19472
1.27464
1.29616
1.30794
1.48334
1.32668
1.26513
1.22270
INA

1.17554
1.16720
1.04214
1.14636
1.15886
1.17137
1.25474
1.19221
1.50068
1.60073
1.35062
1.19221
1.37980
INA

1.14243
1.27712
1.21358
1.19810
1.26454
1.18345
1.27543
1.11649
1.76653
1.49981
1.29400
1.19335
1.28030
INA

1.22717
1.20308
1.35425
1.29945
1.13181
1.12428
1.11088
1.11198
1.62120
1.42244
1.30100
1.14592
1.31618
INA

1.27319
1.12977
1.26023
1.10479
1.07179
1.07782
1.04321
0.91347
1.59645
1.37161
1.11197
0.98238
1.20652
INA

1.08053
1.03515
1.20606
1.18407
0.98807
0.86584
1.01185
0.99336
1.35236
1.25468
0.94397
0.90564
0.98229
INA

0.97181
0.90660
1.31255
1.10233
0.90740
0.79778
1.00230
0.93301
1.06763
1.15280
1.06865
1.08441
1.02225
INA

HAWAII

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

1.85068
1.49146
0.84566
0.75321
0.87707
1.23922
1.07589
1.01368
1.94181
0.71217
0.64151
0.84177
1.49766
0.65360

1.45962
1.45747
0.82988
0.70916
0.63358
1.13715
1.09110
1.06033
1.67458
0.75929
0.38345
0.75643
1.31674
INA

0.96957
0.93930
1.01324
0.74242
0.82243
0.83216
1.21538
1.35436
1.88946
1.07004
0.53083
0.92263
1.37951
INA

1.00000
0.82615
0.96750
0.68352
1.01121
0.82681
1.32032
0.92119
1.23347
1.27745
0.72236
1.02679
1.13351
INA

1.34378
0.38973
1.00672
1.40860
1.20233
1.04397
0.80000
1.17621
0.96714
1.54123
0.98032
0.87099
0.83302
INA

1.44719
1.34719
1.40622
1.11209
1.28796
1.03875
1.16286
1.01138
0.89344
1.01898
1.24628
0.88309
0.97254
INA

1.30420
1.30420
1.30420
1.13031
1.13031
0.95641
0.95641
0.95641
1.04336
1.13031
1.04336
0.95641
0.95641
INA

1.73669
1.29460
1.08151
1.13896
1.15384
0.84950
0.79333
1.52877
1.51738
1.17621
0.91154
1.10322
1.06775
INA

1.27328
1.31928
1.02552
0.68622
0.88447
0.97578
0.92254
1.44682
1.67618
1.24560
0.67997
0.82516
1.12769
INA

1.24447
1.29896
1.07227
0.93079
0.97473
0.61880
1.15899
1.19698
1.72664
1.24730
0.67077
0.97360
1.21667
INA

1.12613
0.99033
0.65358
0.72116
1.02276
0.60985
1.13340
1.18237
1.37530
1.15058
0.63568
0.84522
1.02827
INA

1.57395
1.01400
0.76932
0.66973
1.11910
0.86687
1.11538
1.58164
1.16206
1.04110
0.75390
1.18135
1.23755
INA

Revised 2/24/03

Monthly Agricultural Factors

LAUS Program Manual 7-33

CALIFORNIA

REGION

YEAR

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

MICHIGAN

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

1.19156
0.97526
0.74215
0.98246
1.16854
0.90967
0.87692
0.77680
0.81982
0.81239
0.75579
0.59330
0.71890
0.59444

1.09421
0.87075
0.70925
0.95379
1.17186
0.90314
0.96542
0.80327
0.91540
0.82636
0.90527
0.50078
0.77559
INA

1.12142
0.83386
0.62788
1.00755
1.13283
1.01930
0.98193
0.70500
0.82907
0.71280
0.98770
0.59826
0.81576
INA

1.00000
0.87897
0.81500
1.10204
1.07688
1.19750
1.00633
0.82355
0.92083
0.81981
1.06595
0.93331
0.80537
INA

1.10103
1.00741
0.97284
1.18198
1.28423
1.15236
1.11925
0.91615
1.08535
0.96241
1.15027
0.98765
0.84773
INA

1.08062
1.09576
1.06486
1.28706
1.39078
1.25137
1.02929
0.90005
1.02504
0.96058
1.09577
1.06265
0.84823
INA

1.16534
1.14287
0.98281
1.28421
1.34973
1.21869
1.08764
0.99591
1.15315
1.21868
1.15316
1.08764
0.92944
INA

1.25954
1.05855
0.99328
1.19805
1.56243
1.16555
1.10056
1.11478
0.98954
1.17067
1.00866
1.15595
0.87848
INA

1.19026
0.94712
0.89634
1.20985
1.53233
1.07338
1.06441
1.09414
0.90046
0.97056
1.01189
1.03519
1.02941
INA

1.19635
1.03234
0.97486
1.21950
1.43775
1.02637
1.01405
1.11002
0.98604
1.00063
1.02716
0.95857
1.13201
INA

1.06413
0.86683
0.95500
1.17920
1.39619
1.00288
0.96051
1.06917
0.86844
0.85234
0.85882
0.77224
0.95141
INA

1.04160
0.78851
0.87796
1.11627
0.96993
0.88842
0.82807
1.04272
0.87845
0.72598
0.67686
0.75548
0.83350
INA

MINNESOTA

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

1.19156
0.97526
0.74215
0.86025
0.85195
0.65705
0.71920
0.66476
0.64142
0.61947
0.61792
0.44719
0.56563
0.48082

1.09421
0.87075
0.70925
0.81121
0.83894
0.66058
0.79594
0.68653
0.71339
0.63122
0.73890
0.36215
0.61278
INA

1.12142
0.83386
0.62788
0.84460
0.79164
0.74633
0.81883
0.60049
0.63056
0.54689
0.80856
0.43551
0.64761
INA

1.00000
0.87897
0.81500
0.91873
0.73064
0.87366
0.84762
0.70185
0.69927
0.62874
0.87502
0.70695
0.64468
INA

1.10103
1.00741
0.97284
0.97829
0.88283
0.85132
0.94264
0.78078
0.82990
0.73741
0.94612
0.74434
0.68113
INA

1.08062
1.09576
1.06486
1.06300
0.95340
0.92558
0.88589
0.76575
0.76921
0.73727
0.91096
0.79896
0.68634
INA

1.16534
1.14287
0.98281
1.03979
0.90447
0.91159
0.94008
0.84750
0.86886
0.93296
0.96144
0.81188
0.75179
INA

1.25954
1.05855
0.97291
0.95433
1.05492
0.88238
0.95013
0.93929
0.74681
0.90276
0.83305
0.86770
0.71058
INA

1.19026
0.94712
0.85560
0.94818
1.04266
0.82399
0.91779
0.91235
0.68092
0.75612
0.82784
0.78240
0.83265
INA

1.19635
1.03234
0.91376
0.94030
0.98719
0.79936
0.87314
0.91649
0.74663
0.78570
0.83265
0.73004
0.91564
INA

1.06413
0.86683
0.87352
0.89197
0.96725
0.79234
0.82576
0.87236
0.65925
0.67872
0.68439
0.59579
0.76957
INA

1.04160
0.78851
0.77612
0.82533
0.68952
0.71726
0.71018
0.84048
0.66802
0.58854
0.52477
0.58810
0.67419
INA

Revised 2/24/03

Monthly Agricultural Factors

7-34 LAUS Program Manual

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

TABLE 7-2 CPS AGRICULTURAL EMPLOYMENT MONTHLY ESTIMATION FACTOR
BY AGRICULTURAL ESTIMATING REGIONS

REGION
WISCONSIN

YEAR
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

JAN
1.19156
0.97526
0.74215
0.84970
0.90812
0.74996
0.74703
0.67127
0.71059
0.57373
0.47937
0.39695
0.41850
0.31658

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT

NOV

DEC

1.09421
0.87075
0.70925
0.79890
0.90761
0.75392
0.82232
0.69551
0.79324
0.56435
0.57463
0.34285
0.44578
INA

1.12142
0.83386
0.62788
0.83053
0.87352
0.85178
0.83617
0.61362
0.71733
0.44469
0.62608
0.40813
0.46195
INA

1.00000
0.87897
0.81500
0.90290
0.82600
0.99712
0.85675
0.71622
0.79665
0.51576
0.67480
0.62268
0.44410
INA

1.10103
1.00741
0.97284
0.96071
0.98734
0.97153
0.95290
0.79673
0.93938
0.61761
0.72749
0.66085
0.46174
INA

1.08062
1.09576
1.06486
1.04366
1.06873
1.05626
0.87587
0.78475
0.88616
0.59457
0.68947
0.71200
0.45122
INA

1.16534
1.14287
0.98281
1.01869
1.03303
1.04020
0.92543
0.86803
0.99716
0.79629
0.72455
0.73173
0.49499
INA

1.25954
1.05855
0.97115
0.94721
1.20478
0.99460
0.93814
0.97098
0.83422
0.76251
0.63779
0.76682
0.46785
INA

1.19026
0.94712
0.85208
0.95342
1.19070
0.91569
0.90911
0.95233
0.73573
0.62935
0.64385
0.67473
0.54823
INA

1.19635
1.03234
0.90848
0.95795
1.12727
0.87532
0.86797
0.96551
0.78827
0.64659
0.65747
0.61232
0.60287
INA

1.06413
0.86683
0.86649
0.92284
1.10441
0.85503
0.82414
0.92924
0.66512
0.54728
0.55573
0.47611
0.50669
INA

1.04160
0.78851
0.76733
0.86979
0.78713
0.75709
0.71317
0.90553
0.65231
0.46231
0.44543
0.45397
0.44390
INA

Revised 2/24/03

Monthly Agricultural Factors

LAUS Program Manual 7-35

Table 7-3 United States Survival Rates by Age, 1990 through 1996
1-Year
Survival
Age Group

1991
Rate

2-Year
Survival
Age
Group

1992
Rate

3-Year
Survival
Age
Group

1993
Rate

4-Year
Survival
Age
Group

1994
Rate

5-Year
Survival
Age
Group

1995
Rate

6-Year
Survival
Age
Group

1996
Rate

7-Year
Survival
Age Group

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85+

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21-25
26-30
31-35
36-40
41-45
46-50
51-55
56-60
61-65
66-70
71-75
76-80
81-85
86+

100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.6
99.4
99.1
98.6
97.9
96.8
95.2
92.6
83.6

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22-26
27-31
32-36
37-41
42-46
47-51
52-56
57-61
62-66
67-71
72-76
77-81
82-86
87+

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.8
99.8
99.8
99.7
99.7
99.6
99.5
99.2
98.8
98.1
97.0
95.6
93.4
90.2
84.8
69.2

3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23-27
28-32
33-37
38-42
43-47
48-52
53-57
58-62
63-67
68-72
73-77
78-82
83-87
88+

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.7
99.7
99.7
99.7
99.6
99.5
99.3
99.1
98.7
98.1
97.1
95.6
93.5
90.1
85.2
76.4
56.6

4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24-28
29-33
34-38
39-43
44-48
49-53
54-58
59-63
64-68
69-73
74-78
79-83
84-88
89+

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.7
99.6
99.6
99.6
99.6
99.6
99.5
99.3
99.1
98.8
98.3
97.4
95.9
93.7
90.7
86.4
80.0
68.7
45.7

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85-89
90+

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.6
99.6
99.5
99.5
99.5
99.5
99.4
99.3
99.1
98.8
98.5
97.9
96.7
94.7
92.0
88.3
82.9
74.9
60.2
36.4

6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26-30
31-35
36-40
41-45
46-50
51-55
56-60
61-65
66-70
71-75
76-80
81-85
86-90
91+

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.6
99.5
99.5
99.4
99.4
99.4
99.4
99.3
99.2
98.9
98.6
98.1
97.3
95.8
93.5
90.3
85.7
79.2
69.7
52.3
28.6

7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27-31
32-36
37-41
42-46
47-51
52-56
57-61
62-66
67-71
72-76
77-81
82-86
87-91
92+

Source: Bureau of Labor Statistics, Office of Employment and Unemployment Statistics. Derived from 1994 (revised) and 1995 (provisional) United States Life Tables (for age groups up to and including 85), and from 1979-1981 Decennial United
States Life Tables (for age groups over 85), all provided by the Department of Health and Human Services.

Survival Rates

7-36 LAUS Program Manual

1990
Age
Group

Table 7-3 United States Survival Rates by Age, 1997 through 2003
8-Year
Survival
Age Group

1998
Rate

9-Year
Survival
Age Group

1999
Rate

10-Year
Survival
Age Group

2000
Rate

11-Year
Survival Age
Group

2001
Rate

12-Year
Survival
Age Group

2002
Rate

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.8
99.7
99.6
99.5
99.4
99.4
99.3
99.3
99.3
99.3
99.3
99.1
98.8
98.3
97.8
96.7
95.0
92.2
88.4
82.9
75.3
64.0
44.8
22.1

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28-32
33-37
38-42
43-47
48-52
53-57
58-62
63-67
68-72
73-77
78-82
83-87
88-92
93+

99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.7
99.6
99.6
99.5
99.4
99.4
99.3
99.3
99.3
99.2
99.2
99.1
98.8
98.3
97.5
96.3
94.3
91.0
86.6
80.2
71.6
58.7
40.1
17.7

9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29-33
34-38
39-43
44-48
49-53
54-58
59-63
64-68
69-73
74-78
79-83
84-88
89-93
94+

99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.7
99.6
99.6
99.5
99.4
99.3
99.3
99.2
99.2
99.2
99.2
99.1
99.0
98.6
98.0
97.1
95.7
93.3
89.7
84.6
77.4
67.6
52.7
33.6
13.2

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85-89
90-94
95+

99.9
99.9
99.9
99.9
99.9
99.9
99.8
99.7
99.7
99.6
99.5
99.5
99.4
99.3
99.2
99.2
99.1
99.1
99.1
99.1
99.0
98.8
98.4
99.7
96.7
95.0
92.2
88.1
82.2
74.2
63.2
46.9
27.8
9.7

11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31-35
36-40
41-45
46-50
51-55
56-60
61-65
66-70
71-75
76-80
81-85
86-90
91-95
96+

99.9
99.9
99.9
99.9
99.8
99.8
99.7
99.7
99.6
99.5
99.4
99.4
99.3
99.2
99.1
99.1
99.0
99.0
99.0
99.0
98.9
98.6
98.2
97.3
96.1
94.2
91.0
86.4
79.7
70.9
58.7
41.0
22.6
7.0

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32-36
37-41
42-46
47-51
52-56
57-61
62-66
67-71
72-76
77-81
82-86
87-91
91-96
97+

99.9
99.9
99.9
99.9
99.7
99.7
99.7
99.6
99.5
99.5
99.4
99.3
99.2
99.1
99.0
99.0
98.9
98.9
98.9
98.8
98.8
98.5
97.9
97.0
95.5
93.3
89.9
84.8
77.7
67.8
54.1
35.5
18.0
5.0

13-Year
2003 Rate Survival Age
Group
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33-37
38-42
43-47
48-52
53-57
58-62
63-67
68-72
73-77
78-82
83-87
88-92
93-97
98+

Source: Bureau of Labor Statistics, Office of Employment and Unemployment Statistics. Derived from 1994 (revised) and 1995 (provisional) United States Life Tables (for age groups up to and including 85),
and from 1979-1981 Decennial United States Life Tables (for age groups over 85), all provided by the Department of Health and Human Services.

99.9
99.9
99.9
99.7
99.7
99.6
99.6
99.5
99.4
99.4
99.3
99.2
99.1
99.0
99.0
98.9
98.8
98.8
98.8
98.7
98.6
98.3
97.6
96.5
94.9
92.4
88.6
82.9
75.0
64.2
49.2
30.3
14.1
3.5

Survival Rates

LAUS Program Manual 7-37

1997
Rate

Seasonal A' Factors

Table 7-4 Seasonal Monthly A ' Factors by Year1

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

1990

1991

1992

1993

1994

1995

1996

0.9686
0.9967
0.9505
0.9011
1.0161
1.2405
1.0940
1.0160
1.0069
0.9796
0.9474
0.8825

0.9685
0.9891
0.9523
0.9121
1.0227
1.1910
1.0981
1.0178
1.0077
0.9940
0.9488
0.8979

0.9892
0.9827
0.9591
0.9112
1.0117
1.1711
1.1005
1.0147
1.0159
0.9952
0.9418
0.9060

1.0036
0.9746
0.9532
0.8933
1.0468
1.1984
1.1152
1.0117
1.0127
0.9907
0.9282
0.8716

1.0284
0.9960
0.9791
0.9226
1.0409
1.1993
1.1022
1.0098
0.9925
0.9706
0.9006
0.8581

1.0307
0.9953
0.9825
0.9351
1.0330
1.2230
1.1161
1.0128
0.9945
0.9542
0.8915
0.8313

1.0052
0.9789
0.9873
0.9236
1.0354
1.2071
1.1443
1.0305
0.9890
0.9595
0.9036
0.8356

1997

1998

1999

2000

2001

2002

0.9947
0.9906
0.9948
0.9266
1.0226
1.2056
1.1628
1.0228
0.9961
0.9608
0.9067
0.8158

0.9904
1.0047
1.0087
0.9131
1.0110
1.2109
1.1791
1.0041
0.9903
0.9667
0.9145
0.8064

0.9830
1.0255
1.0066
0.9116
0.9951
1.2155
1.1574
0.9922
0.9988
0.9786
0.9172
0.8188

0.9839
1.0279
1.0012
0.9031
0.9945
1.2176
1.1259
0.9991
1.0059
0.9764
0.9358
0.8286

0.9888
1.0046
0.9830
0.9168
0.9836
1.2211
1.1130
1.0002
1.0086
0.9649
0.9461
0.8694

0.9883
1.0022
0.9915
0.9332
1.0021
1.1834
1.0923
1.0060
0.9930
0.9540
0.9652
0.8888

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

Revised 02/24/03

Note: 1Year refers to the last year of CPS data used to develop the factors. These factors
are used in the benchmark revision of that year's Handbook estimates when possible and in
the following year's preliminary and revised estimates.

7-38 LAUS Program Manual

Seasonal B' Factors

Table 7-5 Seasonal Monthly B ' Factors by Year 1

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

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

1990

1991

1992

1993

1994

1995

1996

0.8064
0.8623
0.8781
0.8829
1.0856
1.3685
1.1345
1.0439
1.0577
1.0609
0.9496
0.8695

0.8014
0.8499
0.8445
0.8984
1.0863
1.3155
1.1485
1.0581
1.0697
1.0904
0.9549
0.8823

0.8257
0.8235
0.8660
0.8906
1.0601
1.2648
1.1523
1.0632
1.0941
1.1027
0.9595
0.8974

0.8578
0.8214
0.8442
0.8745
1.1013
1.2611
1.1629
1.0635
1.0924
1.0953
0.9567
0.8690

0.8812
0.8471
0.8804
0.9186
1.1086
1.2695
1.1445
1.0490
1.0600
1.0609
0.9307
0.8498

0.8854
0.8608
0.8910
0.9329
1.1011
1.3067
1.1362
1.0401
1.0568
1.0378
0.9198
0.8313

0.8632
0.8577
0.9033
0.9187
1.1196
1.2954
1.1359
1.0460
1.0534
1.0456
0.9266
0.8345

1997

1998

1999

2000

2001

2002

0.8203
0.8682
0.9080
0.9259
1.1226
1.2933
1.1607
1.0377
1.0640
1.0642
0.9256
0.8095

0.8109
0.8839
0.9235
0.9145
1.1200
1.2952
1.1610
1.0191
1.0501
1.0710
0.9575
0.7932

0.8116
0.9055
0.9285
0.9242
1.1135
1.2833
1.1284
0.9947
1.0501
1.0876
0.9592
0.8135

0.8187
0.8998
0.9176
0.9318
1.1117
1.2843
1.0840
0.9917
1.0477
1.0853
0.9924
0.8350

0.8231
0.8713
0.8891
0.9456
1.0929
1.2766
1.0732
0.9904
1.0522
1.0781
1.0061
0.9014

0.8221
0.8694
0.8953
0.9626
1.1087
1.2440
1.0764
1.0115
1.0539
1.0461
1.0092
0.9009
Revised 02/24/03

Note: 1 Year refers to the last year of CPS data used to develop the factors. These factors
are used in the benchmark revision of that year's Handbook estimates when possible and in
the following year's preliminary and revised estimates.

LAUS Program Manual 7-39

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate

Months for Selecting

Unemployment

Weekly

Use Period

Survival Rates

January-March
1987

April-June
1986

April-June
1987

July-September
1986

July-September
1987

October-December
1986

October-December
1987

January-March
1987

January-March
1988

April-June
1987

April-June
1988

July-September
1987

July-September
1988

October-December
1987

October-December
1988

January-March
1988

January- March
1989

April-June
1988

April-June
1989

July-September
1988

Rate Range
0.0 - 6.4
6.5 - 7.8
7.9 - 9.4
9.5 and up
0.0 - 6.3
6.4 - 7.7
7.8 - 9.0
9.1 and up
0.0 - 5.6
5.7 - 7.1
7.2 - 8.7
8.8 and up
0.0 - 6.3
6.4 - 7.9
8.0 - 9.0
9.1 and up
0.0 - 5.5
5.6 - 7.5
7.6 - 8.7
8.8 and up
0.0 - 5.4
5.5 - 6.5
6.6 - 8.0
8.1 and up
0.0 - 5.1
5.2 - 5.9
6.0 - 7.3
7.4 and up
0.0 - 5.6
5.7 - 7.6
7.7 - 8.5
8.6 and up
0.0 - 5.1
5.2 - 6.1
6.2 - 7.4
7.5 and up
0.0 - 4.8
4.9 - 5.5
5.6 - 7.0
7.1 and up

Survival Rates
0.950
0.956
0.960
0.960
0.946
0.946
0.955
0.955
0.941
0.949
0.953
0.956
0.954
0.960
0.961
0.964
0.960
0.961
0.961
0.964
0.954
0.960
0.961
0.964
0.946
0.946
0.955
0.960
0.939
0.944
0.954
0.954
0.939
0.944
0.954
0.954
0.948
0.955
0.956
0.966
Revised 02/24/03
LAUS Prorgam Manual 7-40

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate

Months for Selecting

Unemployment

Weekly

Use Period

Survival Rates

July-September
1989

October-December
1988

October-December
1989

January-March
1989

January-March
1990

April-June
1989

April-June
1990

July-September
1989

July-September
1990

October-December
1989

October-December
1990

January-March
1990

January-March
1991

April-June
1990

April-June
1991

July-September
1990

July-September
1991

October-December
1990

October-December
1991

January-March
1991

Rate Range
0.0 - 4.3
4.4 - 5.2
5.3 - 6.4
6.5 and up
0.0 - 5.1
5.2 - 5.7
5.8 - 7.0
7.1 and up
0.0 - 4.3
4.4 - 5.3
5.4 - 6.6
6.7 and up
0.0 - 4.8
4.9 - 5.2
5.3 - 6.8
6.9 and up
0.0 - 4.8
4.9 - 5.5
5.6 - 6.2
6.3 and up
0.0 - 5.3
5.4 - 5.8
5.9 - 6.6
6.7 and up
0.0 - 5.0
5.1 - 5.3
5.4 - 6.1
6.2 and up
0.0 - 5.0
5.1 - 5.7
5.8 - 6.4
6.5 and up
0.0 - 5.3
5.4 - 5.8
5.9 - 6.1
6.2 and up
0.0 - 6.8
6.9 - 7.3
7.4 - 7.6
7.7 and up

Survival Rates
0.944
0.944
0.955
0.957
0.942
0.942
0.946
0.950
0.939
0.946
0.947
0.950
0.939
0.946
0.947
0.950
0.937
0.944
0.944
0.947
0.943
0.947
0.954
0.959
0.943
0.944
0.944
0.947
0.931
0.942
0.944
0.944
0.935
0.939
0.944
0.944
0.947
0.951
0.953
0.960
Revised 02/24/03

LAUS Prorgam Manual 7-41

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate

Months for Selecting

Unemployment

Weekly

Use Period

Survival Rates

January-March
1992

April-June
1991

April-June
1992

July-September
1991

July-September
1992

October-December
1991

October-December
1992

January-March
1992

January-March
1993

April-June
1992

April-June
1993

July-September
1992

July-September
1993

October-December
1992

October-December
1993

January-March
1993

January-March
1994

April-June
1993

April-June
1994

July-September
1993

Rate Range
0.0 - 6.2
6.3 - 6.9
7.0 - 7.6
7.7 and up
0.0 - 6.4
6.5 - 7.0
7.1 - 7.6
7.7 and up
0.0 - 6.3
6.4 - 7.0
7.1 - 7.4
7.5 and up
0.0 - 7.7
7.8 - 8.4
8.5 - 8.8
8.9 and up
0.0 - 6.5
6.6 - 7.9
8.0 - 8.8
8.9 and up
0.0 - 6.8
6.9 - 8.4
8.5 - 9.0
9.1 and up
0.0 - 6.3
6.4 - 7.1
7.2 - 8.0
8.1 and up
0.0 - 7.2
7.3 - 7.8
7.9 - 8.6
8.7 and up
0.0 - 6.8
6.9 - 7.2
7.3 - 8.1
8.2 and up
0.0 - 6.6
6.7 - 7.2
7.3 - 7.8
7.9 and up

Survival Rates
0.946
0.950
0.950
0.951
0.946
0.952
0.952
0.957
0.952
0.959
0.959
0.962
0.965
0.966
0.966
0.970
0.955
0.963
0.963
0.967
0.951
0.959
0.960
0.961
9.490
0.961
0.961
0.961
0.960
0.964
0.967
0.967
0.959
0.961
0.963
0.969
0.950
0.952
0.960
0.961
Revised 02/24/03
LAUS Prorgam Manual 7-42

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate
Use Period

Months for Selecting
Survival Rates

July-September
1994

October-December
1993

October-December
1994

January-March
1994

January-March
1995

April-June
1994

April-June
1995

July-September
1994

July-September
1995

October-December
1994

October-December
1995

January-March
1995

January-March
1996

April-June
1995

April-June
1996

July-September
1995

July-September
1996

October-December
1995

October-December
1996

January-March
1996

LAUS Prorgam Manual 7-43

Unemployment
Rate Range
0.0 - 6.0
6.1 - 6.6
6.7 - 7.4
7.5 and up
0.0 - 6.1
6.2 - 7.1
7.2 - 8.2
8.3 and up
0.0 - 5.7
5.8 - 6.6
6.7 - 7.8
7.9 and up
0.0 - 5.4
5.5 - 6.3
6.4 - 8.1
8.2 and up
0.0 - 4.8
4.9 - 5.7
5.8 - 7.0
7.1 and up
0.0-5.4
5.5-6.3
6.4-7.3
7.4 and up
0.0-5.2
5.3-6.0
6.1-6.8
6.9 and up
0.0-5.2
5.3-6.0
6.1-6.7
6.8 and up
0.0-4.7
4.8-5.8
5.9-6.3
6.4 and up
0.0-5.0
5.1-5.6
5.7-6.7
6.8 and up

Weekly
Survival Rates
0.954
0.961
0.961
0.964
0.959
0.959
0.960
0.967
0.955
0.956
0.956
0.960
0.950
0.951
0.957
0.960
0.950
0.955
0.955
0.955
0.951
0.958
0.964
0.964
0.953
0.953
0.958
0.959
0.944
0.948
0.955
0.955
0.946
0.956
0.957
0.964
0.956
0.956
0.957
0.971
Revised 02/24/03

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate
Use Period

Months for Selecting
Survival Rates

January-March
1997

April-June
1996

April-June
1997

July-September
1996

July-September
1997

October-December
1996

October-December
1997

January-March
1997

January-March
1998

April-June
1997

April-June
1998

July-September
1997

July-September
1998

October-December
1997

October-December
1998

January-March
1998

January-March
1999

April-June
1998

April-June
1999

July-September
1998

Unemployment
Rate Range
0.0-4.6
4.7-5.1
5.2-6.0
6.1 and up
0.0-4.6
4.7-5.6
5.7-6.8
6.9 and up
0.0-4.5
4.6-5.1
5.2-6.2
6.3 and up
0.0-5.1
5.2-5.7
5.8-6.8
6.9 and up
0.0-4.6
4.7-5.1
5.2-6.1
6.2 and up
0.0-4.2
4.3-5.1
5.2-6.2
6.3 andup
0.0-4.3
4.4-4.6
4.7-5.7
5.8 and up
0.0-4.0
4.1-4.8
4.9-5.9
6.0 and up
0.0-4.1
4.2-4.8
4.9-5.3
5.4 and up
0.0-4.2
4.3-4.6
4.7-6.0
6.1 and up

Weekly
Survival Rates
0.946
0.953
0.959
0.959
0.944
0.947
0.956
0.956
0.941
0.947
0.952
0.952
0.951
0.955
0.958
0.962
0.950
0.950
0.952
0.962
0.943
0.950
0.957
0.957
0.945
0.953
0.953
0.958
0.947
0.947
0.959
0.966
0.947
0.957
0.961
0.961
0.935
0.946
0.948
0.948
Revised 02/24/03
LAUS Prorgam Manual 7-44

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate
Use Period

Months for Selecting
Survival Rates

July-September
1999

October-December
1998

October-December
1999

January-March
1999

January-March
2000

April-June
1999

April-June
2000

July-September
1999

July-September
2000

October-December
1999

October-December
2000

January-March
2000

January-March
2001

April-June
2000

April-June
2001

July-September
2000

July-September
2001

October-December
2000

October-December
2001

January-March
2001

LAUS Prorgam Manual 7-45

Unemployment
Rate Range
0.0-3.6
3.7-4.4
4.5-5.5
5.6 and up
0.0-4.2
4.3-4.8
4.9-6.1
6.2 and up
0.0-4.0
4.1-4.5
4.6-4.8
4.9 and up
0.0-4.2
4.3-4.7
4.8-5.0
5.1 and up
0.0-3.8
3.9-4.0
4.1-4.6
4.7 and up
0.0-4.3
4.4-4.6
4.7-4.9
5.0 and up
0.0-3.8
3.9-4.1
4.2-4.8
4.9 and up
0.0-3.8
3.9-4.3
4.4-4.9
5.0 and up
0.0-3.5
3.6-4.0
4.1-4.5
4.6 and up
0.0-3.5
3.6-4.0
4.1-4.5
4.6 and up

Weekly
Survival Rates
0.940
0.945
0.946
0.946
0.951
0.953
0.953
0.953
0.946
0.952
0.952
0.955
0.945
0.947
0.947
0.947
0.940
0.946
0.946
0.950
0.951
0.951
0.951
0.961
0.943
0.943
0.943
0.951
0.939
0.942
0.947
0.947
0.938
0.938
0.938
0.943
0.941
0.945
0.945
0.947
Revised 02/24/03

Quarterly Survival Rates
Table 7-6 Quarterly Average Exhaustee Survival Rates Based on Unemployment Rates
Survival Rate
Use Period

Months for Selecting
Survival Rates

January-March
2002

April-June
2001

April-June
2002

July-September
2001

July-September
2002

October-December
2001

October-December
2002

January-March
2002

January-March
2003

April-June
2002

Unemployment
Rate Range
00.-3.2
3.3-3.9
4.0-4.4
4.5 and up
0.0-4.2
4.3-5.0
5.1-5.7
5.8 and up
0.0-5.2
5.3-5.5
5.6-5.8
5.9 and up
0.0-5.5
5.6-6.3
6.4-6.8
6.9 and up
0.0-5.3
5.4-5.9
6.0-6.3
4.4 and up

Weekly
Survival Rates
0.945
0.945
0.945
0.946
0.942
0.943
0.943
0.943
0.951
0.953
0.953
0.955
0.957
0.959
0.960
0.960
0.956
0.961
0.962
0.965
Revised 02/24/03

LAUS Prorgam Manual 7-46

Additivity

8

LAUS Estimation: Additivity

inking substate labor force estimates to the CPS begins with a set of
independent Handbook employment and unemployment estimates. These
are prepared for all LMAs—that is, officially-designated Metropolitan
Areas (MAs), single-county labor market areas, multi-county areas, and
aggregations of cities and town in New England—such that they exhaust all
geographic subdivisions of a State. Because of nonlinearity in the Handbook, the
LMA 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 LMA estimates additive to State levels.

L

Usual statistical practice is to distribute aggregation differences proportionally
among the individual parts. In this manner, all components receive a
proportionate share of the difference between the sum of the parts and an
independent total. The LAUS program uses this simple linear additivity
adjustment method, referred to as the Handbook-Share technique, to adjust LMA
estimates to the State control totals. 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 assumes a proportional
distribution throughout the State of the difference between the sum of the
independent Handbook estimates and the State control totals. This method is
applied to all areas for which an independent Handbook estimate is prepared and
to the intrastate portions of interstate areas. The adjustments for additivity are
performed on a current basis, and whenever the State estimates are revised.
After the Handbook estimates have been adjusted for additivity to the statewide
estimates, LMA estimates are referred to as “LAUS” estimates rather than
Handbook estimates. These LAUS estimates are then disaggregated further into
smaller areas, such as single counties within multi-county LMAs, or sub-county
areas, such as cities and towns, for which estimates may be required by
legislation. Two methods for disaggregation exist based on the availability of UI

LAUS Program Manual 8-1

claims and decennial census data for apportioning LMA estimates to
smaller areas. As with the requirement for additivity of LMA estimates to
statewide totals, sub-LMA estimates produced by disaggregation are
additive to the LMA estimates. See Chapter 9 for a complete description
of the disaggregation process.

8-2 LAUS Program Manual

Adjustment to Independent Statewide Estimates—The Handbook Share Method

Adjustment to Independent Statewide Estimates—The
Handbook Share Method
The process of reconciling, or linking,
LMA labor force estimates to Statewide
S
(model-based) estimates begins with a set of
L
geographically exhaustive LMA Handbook
P
M
employment and unemployment estimates.
A simultaneous adjustment for additivity of
A C
all LMA estimates to the statewide
estimates is performed using the percentage
distribution of the substate Handbook
estimates, also known as the HandbookShare 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 Handbook-Share method should be followed by all States unless the State can
demonstrate and document why the linear additivity adjustment procedure is
inappropriate. The State must also be able to suggest a reasonable and equitable
alternative distribution. Reasons for alternative procedures may include
inconsistent quality of employment estimates or deficiencies in the Handbook
estimates for minor LMAs. Documentation should show how the alternative
procedure differs from linear adjustment in terms of the distribution of
employment and unemployment in the State. Linear additivity adjustments
should be reviewed annually and exception requests should be submitted to the
Regional Office before annual benchmarking.
The following worksheet illustrates simultaneous additivity and adjustments to
LMAs using the Handbook-Share method.

LAUS Program Manual 8-3

Adjustment to Independent Statewide Estimates—The Handbook Share Method

Simultaneous Additivity of LMA Estimates Using the
Handbook-share Method
Employment

Unemployment

Area

Handbook

Percent of
Summed
Handbook

State

Statewide*

Handbook

Percent of
Summed
Handbook

49,300

Statewide*

562,800

MSA 1

18,500

0.394456

19,447

190,600

0.3481279

195,926

Major LMA 1

9,300

0.198294

9,776

107,100

0.1956164

110,093

Major LMA 2

8,700

0.185501

9,145

103,400

0.1888585

106,290

Minor LMA 1

2,300

0.049041

2,418

36,800

0.0672146

37,828

Minor LMA 2

1,900

0.045120

1,997

25,900

0.0493059

26,624

Intrastate Portion of Interstate MSA 2

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.

8-4 LAUS Program Manual

Interstate Areas

Interstate Areas
For interstate areas, after the independent Handbook estimate is prepared by the
“controlling” State, the intrastate portions are calculated through the
disaggregation process. These intrastate portions are then adjusted for additivity
to the respective statewide levels by each State. The intrastate portions are then
summed to the total interstate area to obtain the employment and unemployment
estimates for the labor market area.

LAUS Program Manual 8-5

Interstate Areas

8-6 LAUS Program Manual

Disaggregation

9

LAUS Estimation:
Disaggregation

Introduction
isaggregation techniques are used to obtain
current estimates of employment and
unemployment for subareas within labor market
areas. Disaggregation involves prorating employment and
unemployment for the labor market area to subarea
jurisdictions. Since these jurisdictions are within LMAs,
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 counties within multicounty labor markets areas, cities within counties (either single-county LMAs or
disaggregated counties), and cities and towns within LMAs in New England.
Two methods of disaggregation are appropriate for LAUS use.
1.) The population-claims method uses current UI claims data by county or
place (city or town) of residence, 1990 census population data by age
group, 1990 census employment data, and the most recent Census
Bureau population estimates. This method uses a separate methodology
for employment (employment/population indexed share) and
unemployment (claims-based unemployment disaggregation).
2.) The census-share method uses 1990 census employment and
unemployment data.

A hierarchy of disaggregation techniques exists. For counties and county
equivalents and cities and towns in New England, the population-claims method
of disaggregation is required since the necessary residency-based claims data are

LAUS Program Manual 9-1

Introduction

available. Outside of New England, for places within counties, such as
cities, the population-claims method of disaggregation is optional,
although preferred. If census data employment and unemployment for the
jurisdiction are not available, contact the BLS Regional Office for
assistance before proceeding.
Note: Throughout this chapter, the term ‘county’ will be used
synonymously with ‘cities or towns’ in New England.
The following tables presents the appropriate method of disaggregation
used for each type of area, based on the availability of claims or published
census data. If a State Employment Security Agency has access to special
tabulations of 1990 census employment and unemployment or age-group
population data for smaller geographical areas that are indicated below,
the preferred disaggregation method, population-claims, may be used
over census-share.

Area

Method

County within multi-county LMA

Population-claims

Incorporated place of 2,500 population or more

Population-claims

Incorporated place of less than 2,500

Census-share

All unincorporated places

Census-share

Within MA, place coterminus with census tracts (not a full county or city)

Census-share

Within MA, place coterminus with
blocks, not with census tracts

Census-share

Outside MA, place coterminus with
Minor Civil Division (MCD) or Census
Division (CD)

Census-share

Outside MA, place not coterminus
with MCD or CD, with population of
1,000 or more

Census-share

The starting point for disaggregation is the estimate of employment and
unemployment prepared for the LMA in accordance with Handbook
instructions outlined in Chapter 7 and the directions on adjustment for
additivity to statewide totals in Chapter 8.

9-2 LAUS Program Manual

The Population-Claims Method of Disaggregation

The Population-Claims Method of Disaggregation
Since current employment and unemployment estimates at the sub-LMA level are
required to implement numerous Federal economic assistance and employment
and training programs, methods of disaggregation which reflect current economic
conditions in these small areas are necessary. Other than the decennial census,
there are very few data series for small areas. Two exceptions are the monthly UI
claims series and the annual 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 for counties and biennially for places
pertaining to July 1 of the given year. The procedure which incorporates the use of
these data is known as the population and claims-based disaggregation procedure.
Population-Based Employment Disaggregation for Counties

Early research showed that, for disaggregating labor market area employment,
annually prepared population estimates alone at the county and city/town level
were superior to the use of fixed decennial census ratios of total employment. The
use of fixed census ratios assumed no change in the ratios over the intercensal
decade. The annual population method allowed the ratios to change over time, but
assumed that the employment/population ratio in each subarea was equal to the
employment/population ratio of the labor market area as a whole. In many
instances, this assumption proved unrealistic.
Subsequent research devised a disaggregation procedure which allows for
differences in employment/population ratios within a LMA. This procedure
utilizes the relationship of the subarea’s employment/population ratio to that of
the larger area, using decennial census data. The assumption is that the
relationship of the employment/population ratios (the ratio of the a subarea’s share
of LMA employment to the subarea’s share of LMA population) will hold
relatively constant in the intercensal period.
This employment disaggregation method can only be used in conjunction with the
claims-based unemployment disaggregation.

LAUS Program Manual 9-3

The Population-Claims Method of Disaggregation

Employment Disaggregation Procedure and
Sequence
A disaggregation procedure fitting the description above can be expressed
as follows:
(Equation 1)
(Eic ÷ Pic)
Pit
×
× Ec = E it
(Ec ÷ Pc)
Pt

Where:
1.) E = total employment
2.) P = total population
3.) c = 1990 census
4.) i = designation of i th county in LMA (variables with no subscript
pertain to the LMA as a whole)
5.) t = reference period of the estimates (For employment, the
reference period is the month. For population, the reference
period is the appropriate year).

When this procedure is used to disaggregate employment, additivity of
the counties to the LMA is not assured, because each county has its own
employment/population index share. Forcing additivity into the
disaggregation yields the following modified formula, known as the
employment/population (E/P) Index share procedure:
(Equation 2)
Eic
h

Eic × Pic

i-1

Pic

∑

×

Pit
Pic

×E=E
t
it

Where E, P, c, i, and t are as defined above and h = the total number of
counties in the LMA. The additivity property can be verified by summing
both the left-hand and right-hand terms.
The Census Bureau prepares total population estimates pertaining to July
1 each year for States and substate areas. Using decennial census data and
the annually-prepared total population estimates, an employment/
population index share is calculated annually for each county in a LMA.

9-4 LAUS Program Manual

The Population-Claims Method of Disaggregation

Given the time lag in issuance of the population data from the Census Bureau,
estimates for the most recent year may not always be available at benchmark
revision time. A LAUS Technical Memorandum, typically issued in late fall each
year, will advise the States of the availability of population estimates.

Applying the Employment Disaggregation Procedure
Each year, data are developed to produce the county employment/population
index-shares, as follows:

Worksheet A.
Developing County Employment/Population Index-Shares
1990 Census

1998

1st Stage
Employment

E/P IndexShare

Employment

Population

Population

I × (III ÷ II)

IV ÷ LMA
Total

I

II

III

IV

V

A

16,500

32,000

35,500

18,305

0.417464

B

12,900

25,300

28,700

14,634

0.333739

C

10,000

22,000

24,000

10,909

0.248797

43,847

1.000000

County

LMA Total
Step 1.

Data from the 1990 census on total employment and population are
entered in Columns I and II for all counties (cities and towns in New
England) in the LMA.

Step 2.

The most recent July 1 population estimates (1998 in this example) are
entered in Column III, and the year of reference indicated.

Step 3.

For each county, a first-stage employment level is calculated by moving
1990 census employment by the change in the county’s population
since the census. Thus, Column IV equals Column I times Column III
divided by Column II.

Step 4.

The county first-stage employment estimates in column IV are
summed to a LMA total (Column IV, LMA Total), Thus, the LMA total
in column IV is a sum and should not be calculated by applying
the formula in the column heading to the LMA total employment
and population.

LAUS Program Manual 9-5

The Population-Claims Method of Disaggregation

Step 5.

The employment/population index-share is calculated for each
county by dividing the first-stage employment level in the
county by the labor market area sum obtained in Step 4. Thus,
Column V equals Column IV divided by the Column IV LMA
Total.

The sum of the employment/population index-shares should equal one,
except for rounding. If it does not precisely equal one, the largest share is
rounded so that the sum of the shares is exactly equal to 1. This will
ensure that employment from the counties sums to the LMA total
employment. The shares are then used in the following Worksheet B,
Employment/Population Index-Share Approach to Disaggregating Total
Employment, as follows:

Worksheet B.
Employment/Population Index-Share Approach to Disaggregating Total Employment
County

E/P IndexShare

January
Employment

February
Employment

I

II

A

0.417464

12,647

12,891

B

0.333739

10,111

C

0.248797

LMA
Total

1.000000

November
Employment

December
Employment

. . .

14,814

15,087

10,306

. . .

11,843

12,061

7,537

7,683

. . .

8,829

8,992

30,295

30,879

. . .

35,485

36,140

. . .
III

Step 1.

From Worksheet A, Column V, enter the employment/
population index-shares in Column II for each county listed in
Column I.

Step 2.

Enter the independent estimate of total employment for the
LMA under the appropriate month in Column III. The
independent estimate for the labor market is the estimate that
results from the application of the Handbook estimating
procedure and adjustment for additivity to the statewide
controls.

Step 3.

The county employment/population index-shares are applied
to the independent LMA employment estimate to arrive at the
disaggregated county employment estimates for the month.
The sum of the disaggregated county employment estimates

9-6 LAUS Program Manual

The Population-Claims Method of Disaggregation

may not add to the LMA total because of rounding. If this is the case,
the estimate for the largest county should be adjusted so that the
summed estimates equal the LMA total.

The same procedures are used to produce employment/population index-shares
for cities:

Worksheet C.
Developing City Employment/Population Index-Shares
1990 Census

1998

1st Stage
Employment

E/P IndexShare

Employment

Population

Population

I × (III ÷ II)

IV ÷ County
Total

I

II

III

IV

V

A

18,300

38,000

42,000

20,226

0.322077

B

14,000

29,500

33,000

15,661

0.249381

Balance of
County

23,600

57,000

65,000

26,912

0.428542

62,800

1.000000

City

County
Total
Step 1.

Data from the 1990 census on total employment and population are
entered in Columns I and II for all LAUS cities in the county. Statespecific cities should not be included. Balance of county employment
and population are derived by subtracting the respective data for LAUS
cities from the county totals.

Step 2.

The most recent July 1 population estimates (1998 in this example) are
entered in Column III, and the year of reference indicated.

Step 3.

For each city and for balance of county, a first-stage employment level
is calculated by moving 1990 census employment by the change in the
city’s or balance of county’s population since the census. Thus, Column
IV equals Column I times Column III divided by Column II.

Step 4.

The city and balance of county first-stage employment estimates are
summed to a county total (Column IV, County Total). Thus, the county
total for column IV is a sum and should not be calculated by
applying the formula in the column heading to the county total
employment and population.

LAUS Program Manual 9-7

The Population-Claims Method of Disaggregation

Step 5.

The employment/population index-share is calculated for each
city by dividing the first-stage employment level in the city by
the county sum obtained in Step 4. Thus, Column V equals
Column IV divided by the Column IV LMA Total.

The sum of the employment/population index-shares should equal one,
except for rounding. If it does not precisely equal one, the largest share is
rounded so that the sum of the shares is exactly equal to 1. This will
ensure that employment from the cities sums to the county total
employment. The shares are then used in the following Worksheet D,
Employment/Population Index-Share Approach to Disaggregating Total
Employment, as follows:

Worksheet D.
Employment/Population Index-Share Approach to Disaggregating Total Employment
City

E/P IndexShare

January
Employment

February
Employment

I

II

A

0.322077

19,582

20,049

B

0.249381

15,162

County
Total

1.000000

60,800

November
Employment

December
Employment

. . .

21,013

21,674

15,524

. . .

16,270

16,782

62,250

. . .

65,243

67,295

. . .
III

Step 1.

From Worksheet C, Column V, enter the employment/
population index-shares in Column II for each city listed in
Column I.

Step 2.

Enter the independent estimate of total employment for the
county under the appropriate month in Column III. The
independent estimate for the county is the estimate that results
from the application of the Handbook estimating procedure and
adjustment for additivity to the statewide controls.

Step 3.

The city employment/population index-shares are applied to
the independent county employment estimate to arrive at the
disaggregated city employment estimates for the month. The
sum of the disaggregated city employment estimates may not
add to the county total because of rounding. If this is the case,
the estimate for the largest city should be adjusted so that the
summed estimates equal the county total.

9-8 LAUS Program Manual

The Population-Claims Method of Disaggregation

Claims-Based Unemployment Disaggregation
Research has shown that the use of current claimant information in disaggregating
labor market area unemployment to subareas is superior to decennial census based
disaggregation because it allows for seasonality during the course of the year and
change during the intercensal period. However, these studies have also shown that
a strict claimant allocation method is not appropriate for total unemployment
because claimants are not representative of the total group of unemployed. This is
particularly true of the entrant-reentrant segment, as these unemployed have a
different seasonal pattern to their joblessness. Disaggregation based solely on
claims data generally underestimates urban areas and inaccurately allocates
blacks, youth, and older women.
In an attempt to correct for this, claims data by county of residence are used to
distribute the experienced unemployed component, i.e., those with recent job
attachment. Decennial census age-group population data aremay not add to the
LMA total because of rounding. If this is the case, the estimate for the largest
county should be adjusted so that the summed estimates equal the LMA total.
Census population data are used in disaggregating unemployed entrants and
reentrants, under the assumption that the population distribution and age structure
of the population within the LMA do not shift drastically over time. Population aged
in 16 to 19 is one element the disaggregation; the other is population aged 20 and over.
Note that these age groups are those used to calculate the youth population ratio for
estimating LMAentrant and reentrant unemployment using the Handbook procedure.
Entrant and reentrant disaggregation ratios are calculated only once every 10
years. In addition, differential migration will have an impact on the LMA’s
distribution of population. However, the lack of current data on migration by age
group at the county level precludes any attempt to correct for this.

Required Claims Data for Claims-Based Unemployment
Disaggregation
For all multi-county LMAs, the residency requirement for claims data is the
coding and tabulating of claimants by county (city or town in New England) of
residence, within the State paying the benefits or in border States if the claimant is
filing under commuter arrangements.
For interstate LMAs, the claims data used in disaggregation must be coded for
residence in counties (cities or towns in New England) in contiguous States where
commuter claimant arrangements exist, as well as within the State paying the
benefits, in order to use the claims-based method to disaggregate the intrastate
portions of the interstate LMA. If commuter claimant data are not available by

LAUS Program Manual 9-9

The Population-Claims Method of Disaggregation

county of residence, the census-share method must be used to estimate
unemployment and employment in each State’s portion of the interstate
LMA. However, in a given State’s intrastate portion, if the State has
claims data by county of residence, the claims-based unemployment
disaggregation (and the employment/population index share method)
must be used to disaggregate to the county level.
The geographic distribution of claimants filing continued claims under
State UI and UCFE certifying to unemployment in the week including the
12th of the month by county of residence is used to disaggregate the LMA
estimate of experienced unemployed to the county level (city or town in
New England). Claimants with any earnings due to employment in the
week including the 12th should be excluded from counts used in
disaggregation.

Unemployment Disaggregation Procedure and
Sequence
The procedure and sequence for unemployment disaggregation is
presented below, along with an example. The example assumes the
following Handbook data:
• unemployment, excluding entrants = 10,000
• B factor unemployment = 900
• A factor unemployment = 1,100
• total unemployment = 12,000
• total LMA claimants without earnings = 6,500
• independent estimate of LMA unemployment = 7,000

LMA Distribution of Population
County

Claimants

> 20 yrs.

16-19 yrs.

1

2,500

25%

20%

2

2,250

30%

35%

3

1,750

45%

45%

Total

6,500

100%

100%

Step 1.

For a LMA, 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.

9-10 LAUS Program Manual

The Population-Claims Method of Disaggregation

Example: 10,000 ÷ 12,000 = .8333
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 LMA Handbook unemployment
represented by new and reentrants to the level of the experienced
unemployed, i.e., B factor unemployment divided by total
unemployment.

Example: 900 ÷ 12,000 = .0750
Step 3.

Determine the proportion of LMA Handbook unemployment
represented by new and reentrants related to the level of the labor
force, i.e., A factor unemployment divided by total unemployment.

Example: 1,100 ÷12,000 = .0917
Note: The proportions obtained in steps 1, 2, and 3 should sum to one (100%).
Example: .8333 + .0750 + .0917 = 1
Step 4.

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

unemployed

b. new and reentrant unemployed related to the experienced

unemployed
c. new

and reentrant unemployment related to the labor
force.
Example:
A = .8333 × 7,000 = 5,833.1
B = .0750 × 7,000 = 525
C = .0917 × 7,000 = 641.9

LAUS Program Manual 9-11

The Population-Claims Method of Disaggregation

Step 5.

Allocate the LMA estimate of experienced unemployed
(estimate A in Step 4) to all counties (cities and towns in New
England) based on the percent distribution of place-ofresidence claims data.

Example:
County 1 = 2,500 ÷ 6,500 = .3846
County 2 = 2,250 ÷ 6,500 = .3462
County 3 = 1,750 ÷ 6,500 = .2692
County 1 = .3846 × 5,833.1 = 2,243.4
County 2 = .3462 × 5,833.1 = 2,019.4
County 3 = .2692 × 5,833.1 = 1,570.3

Step 6.

Allocate the LMA estimate of new and reentrant employment
(estimate B in Step 4) to all counties based on the percent
distribution of the LMA’s population 20 years of age and older
from the 1990 census.

Example:
County 1 = 525 × .25 = 131.25
County 2 = 525 × .30 = 157.50
County 3 = 525 × .45 = 236.25
Step 7.

Allocate the LMA estimate of new and reentrant
unemployment related to the labor force (estimate C in Step 4)
to all counties based on the percent distribution of the LMA’s
population 16-19 years old from the 1990 census.

Example:
County 1 = 641.9 × .20 = 128.38
County 2 = 641.9 × .35 = 224.67
County 3 = 641.9 × .45 = 288.86

9-12 LAUS Program Manual

The Population-Claims Method of Disaggregation

Step 8.

Derive the total unemployment estimate for each county by summing
the county estimates derived in Steps 5, 6, and 7. The sum of the
county unemployment estimates should automatically equal the LMA
total unemployed. If they are not equal due to rounding, the data for the
largest county is adjusted accordingly.

Example:
County 1 = 2,243.4 + 131.25 + 128.38 = 2,503.03
County 2 = 2,019.4 + 157.50 + 224.67 = 2,401.57
County 3 = 1,570.3 + 236.25 + 288.86 = 2,095.41
Total LMA unemployed = 7,001.01

Population-Claims Disaggregation of Interstate Areas
In interstate LMAs where all States have the necessary claims data by county of
residence, a “Handbook equivalent” for each intrastate portion is disaggregated
from the total LMA Handbook estimate using the population-claims method. The
data for each intrastate portion are adjusted for additivity to the respective
statewide controls. Then the adjusted intrastate portion is disaggregated to the
county level by the population-claims method. The entire LMA distribution of
unemployment, excluding entrants, B factor unemployment, A factor
unemployment relative to total Handbook unemployment should be applied to the
intrastate portion of total unemployment to obtain the three groups of unemployed
which are then disaggregated to county (city and town in New England) following
Steps 5 through 8.
In interstate areas where commuter claimant data are not available for all parts, the
census-share method must be used to estimate both employment and
unemployment for the intrastate portions of the interstate LMA. However, in a
given intrastate portion, if the portion is a multi-county area and the State has
claims by county of residence (city and town in New England), the populationclaims method must be used to disaggregate to the county level. In this case, the
census-share total unemployment ratio of the intrastate portion to the whole LMA
should be applied to unemployment, excluding entrants, B factor unemployment,
A factor unemployment, and total unemployment to obtain a Handbook
“equivalent” estimate for the intrastate portion. After this intrastate portion is
adjusted for additivity to the statewide controls, the population-claims method
must be used to disaggregate to the county level.

LAUS Program Manual 9-13

The Population-Claims Method of Disaggregation

Disaggregating Employment and Unemployment to
Incorporated Places Using the Population-Claims
Method
Simple modifications of the employment/population index share
employment disaggregation and the claims-based unemployment
disaggregation enable the development of labor force estimates for units
of local government as small as 2,500 population (according to the 1990
census data), provided claims data are available by residence of the
claimant in all such places in the State. The State may specify the
population level of the places to which this disaggregation method will
apply.
In addition, the balance-of-county estimates (derived after subtracting the
disaggregated place estimates) must relate to a specifically defined
geographic area. Census data for this geographic area must be available
for disaggregation to other places in the balance-of-county area. If such
census data are not available, current claims and population data cannot
be used to disaggregate any place within the county. In places where
commuter claimant arrangements exist, further specification of the claims
data is required.
Place estimates disaggregated by the population-claims method should be
introduced the first month for which residency claims data are available.
Once this method is initiated, it must be used for the rest of the calendar
year. At benchmarking time, the State may opt to return to censussharing, in which case the full time series is revised using the censusshare method.
Specification of Population Size for Place Disaggregation

The place level to which the population-claims method is used is
established based on the last decennial population size of cities and towns
and is adjusted over the intercensal period by the annually prepared
county and city population estimates.
In the intercensal period, the State should review the total population
estimates for all units of local government issued annually by the Census
Bureau to determine whether population changes have occurred which
affect the composition of the size class for place disaggregation. If,
because of a reduction in population, a city falls below the size
specification for place disaggregation using the population-claims
method, the State does not have to revert to the census-share for that

9-14 LAUS Program Manual

The Population-Claims Method of Disaggregation

place, and may continue to use the population-claims. In addition, the collection
of residency-based claims data does not have to be extended to other cities in the
smaller size class.
A city can move into the size class specified for claims-population disaggregation
of places due to an increase in population. If a State is already using the
population-claims method at the place level, it has one year to develop the
residency claims data needed for disaggregation by coding claims by place of
residence for the newly-added city. During that year, the city would continue to be
census-shared. If the residency data are not developed for the newly-added city
after one year, the State cannot use the population-claims method and must revert
to the census-share method for all cities in the size class. Alternately, the State
may avoid reverting to the census-share method by redefining upward the size
class for place disaggregation at the benchmarking time. Then, the populationclaims method can be used in a size class covering larger cities.
If a city, due to a change in population, moves either in or out of the size specified
for population-claims disaggregation of places, the State should notify BLS of this
change and the subsequent methodology changes.
Population-Based Employment Disaggregation

The Census Bureau issues total population estimates for all units of local
government annually. The latest population estimates and the employment/
population index-shares calculated from the census can be used to disaggregate
LMA employment estimates below the county level, using the procedure
described in Worksheet A. Balance-of-county estimates must also be calculated to
allow for the proper application of the additivity adjustment.
The case of disaggregating to a place from a single-county LMA is
straightforward. The employment/population index share procedure, which
involves the use of decennial census employment and population data and
annually prepared population estimates, is applied to the LMA total employment
estimate for the month in question to obtain the place estimate.
In the case of disaggregating to a place from a county in a multi-county LMA, the
county total employment estimate must first be prepared. Population data for
counties and places may not be available on the same time frame. For example,
1995 data may be available at the county level, but for cities the most recent data
may be from 1994. In this case, 1995 data would be used to disaggregate to the
county level, and 1994 data for both the county and cities would be used for
disaggregating from the county level to cities within the county. Thus, the
disaggregated county employment is further broken down to the place (city) by
using the place’s most recent population, the county’s population for that same
year, and the place’s employment/population index-shares from the decennial

LAUS Program Manual 9-15

The Population-Claims Method of Disaggregation

census. The employment/population ratio is then applied to the county
employment estimate for the month in question to obtain the place
employment estimate.
In developing place data using the employment/population index-share
approach, States are reminded that they are to calculate index shares for
all places which meet the chosen population specifications and not just for
those which are reported to BLS. States are then to calculate a rest-ofcounty estimate by subtracting all disaggregated estimates from the
county total. The rest-of-county estimates must relate to a specific
geographic area for which census data exist, so that the census-share
procedure can be used for disaggregation. If this is not possible, the index
share approach cannot be used.
Claims-Based Unemployment Disaggregation

Unemployment in a place of 2,500 population or more may be
disaggregated directly from the intrastate LMA (either single-county or
multi-county) depending on the existence of commuter claims
arrangements and the availability of commuter claims coded by city of
residence and on the same reference period. Disaggregating directly from
the LMA cannot be done for interstate areas because interstate areas must
first be broken down into intrastate portions. Then, unemployment may be
disaggregated directly from those portions, based on the conditions
described above.
Geographic Basis of Claims Data Used to Distribute Experienced
Unemployed

In the claims-based disaggregation, LMA unemployment is disaggregated
into three basic components: the experienced unemployed, unemployed
entrants related to experienced unemployed, and unemployed entrants
related to the labor force. The experienced unemployed component is
distributed to areas based on the distribution of claims. In the case of
place disaggregation, if commuter claims arrangements exist and these
claims are coded and tabulated for city of residence, then the experienced
unemployed distributor is as follows:
Claimants residing in the city who file either in their own State or the
border State, as a percent of all residents of the county (in the case of
single-county areas) or intrastate portion (in the case of multi-county
interstate areas and New England interstate LMAs) who file in the State
or in the border State.

If commuter claimant arrangements exist, but commuter claims are coded
for county of residence only and not city, then the experienced
unemployed distributor is the following:

9-16 LAUS Program Manual

The Population-Claims Method of Disaggregation

Claimants residing in the city who file in the State as a percent of all residents in
the county filing in the State.

That is, commuter claims are not used at all. Use of this modified ratio avoids
distorting the city’s share of the experienced unemployed, while allowing the
county-coded commuter claims to be used.
Unemployment Disaggregation Procedure for Cities or Towns

The following disaggregation is used in almost all cases. It is the same procedure
for claims-based county disaggregation described earlier, with the following
modifications:
Step 5.

Step 6.

Step 7.

Allocations are based on current claims data by city or town of
residence.
Allocations are based on population 20 years of age and older.
Allocations are based on population 16-19 years of age for places with
a 1990 population between 2,500 and 10,000.

Use of these ratios parallels the use of county to multi-county LMA ratios. The
total unemployment estimate for the place (Step 8) is then the sum of the
disaggregated experienced unemployment (Step 5) and the new and reentrants
(Steps 6 and 7).
The procedure above can be used for places in:
1.) Single or multi-county LMAs not contiguous to a border State;
2.) Single or multi-county LMAs contiguous to a border State without
commuter claimant arrangements; and
3.) Single or multi-county LMAs with commuter claimant arrangements
where such claimants are also coded for city of residence.

Modifications to this procedure are required in (1) interstate areas and (2) single
or multi-county LMAs contiguous to a State with commuter claimant
arrangements where such claimants are coded for county of residence only and
not city.
In the case of an interstate area or single- or multi-county LMA contiguous to a
State in which commuter claims are coded for county of residence only, the Step 5
proportion must be based on intrastate claims only, with claims data limited to
residents of the county and the city filing in the State. The proportion becomes the
ratio of city residents filing in the State to county residents filing in the State. The
ratio of city residents filing in the State to county residents filing in the State and

LAUS Program Manual 9-17

The Population-Claims Method of Disaggregation

in the border State will underestimate experienced unemployed in the city.
For multi-county areas (including the intrastate portion of interstate
areas), it is necessary to first disaggregate to the county level before
disaggregating to the place in order to use the county-coded commuter
claimant data. Steps 6 and 7 are on the same geographic reference as Step
5, that is, the city as a percent of the county. Step 8 is the sum of the
disaggregated experienced unemployed (Step 5) and new entrants and
reentrants (Steps 6 and 7).

9-18 LAUS Program Manual

Use of 1990 Census Data in Disaggregating Labor Force Estimates—Census-Share Method

Use of 1990 Census Data in Disaggregating Labor Force
Estimates—Census-Share Method
The use of 1990 census data for disaggregating labor force estimates is required
when more current data for disaggregation are not available. This typically will
occur for administrative areas such as Areas of Substantial Unemployment,
unique geopolitical areas such as Indian reservations, and very small areas such as
parts of cities.
The census-share method uses employment and unemployment ratios. These
ratios are applied to independent single county LMA estimates after adjustment to
State controls, or to disaggregated sub-LMA levels which were based on those
independent LMA estimates.
When the claims-based unemployment disaggregation and population-based
employment disaggregation are used to disaggregate a place in a county, the
balance-of-county area must be a geographic area for which 1990 census data are
available for disaggregating to other places in the balance-of-county area.

The Census-Share Method of Disaggregation
The census-share method of disaggregation utilizes the ratios of employment and
unemployment in a subarea to the respective total for the larger area according to
the 1990 census. These ratios are applied to the current total employment and
unemployment estimates for the larger area. This procedure is based on the
assumption that the current geographic distribution of employment (or
unemployment) is the same as that in the decennial census, or, equivalently, that
employment (unemployment) in the subarea has changed by the same proportion
since the census as that in the larger area. The 1990 census-share procedure is
used to disaggregate from the county to a subcounty area when census labor force
data are available and a State opts not to use the claims and population-based
disaggregation procedure at the city level. If census labor force data are not
available, contact the BLS regional office to make an atypical request to use the
population-share procedure.

LAUS Program Manual 9-19

Use of 1990 Census Data in Disaggregating Labor Force Estimates—Census-Share Method

Disaggregation Procedure and Sequence
The procedure and sequence disaggregation using the census-share
method, along with an example, is presented below. For the example, the
following data are given:

Census data:
LMA employment = 20,000

County 1 = 10,000
County 2 = 6,000
County 3 = 4,000

LMA unemployment = 8,000

County 1 = 4,000
County 2 = 2,400
County 3 = 1,600
•

Independent estimate of LMA unemployment = 7,000

• Independent estimate of LMA employment = 35,000
Step 1.

From the 1990 census data, obtain the number of employed in
the county.

Step 2.

From the 1990 census data, obtain the number of employed in
the LMA containing the county.

Step 3.

Divide Step 1 by Step 2. The result is the ratio of the county
employment to that of the LMA as of April 1990.

Example:
County 1 = 10,000 ÷ 20,000 = .5
County 2 = 6,000 ÷ 20,000 = .3
County 3 = 4,000 ÷ 20,000 = .2

9-20 LAUS Program Manual

Use of 1990 Census Data in Disaggregating Labor Force Estimates—Census-Share Method

Step 4.

Apply the ratio developed in Step 3 to the total employment estimate
for the LMA for the relevant time period. This will yield the estimate of
total employment in the county.

Example: Employment
County 1 = 35,000 × .5 = 17,500
County 2 = 35,000 × .3 = 10,500
County 3 = 35,000 × .2 = 7,000
Step 5.

From the 1990 census data, obtain the number of unemployed in the
county.

Step 6.

From the 1990 census data, obtain the number of unemployed in the
LMA containing the county.

Step 7.

Divide Step 5 by Step 6. The result is the ratio of the county
unemployment to that of the LMA as of April 1990.

Example:
County 1 = 4,000 ÷ 8,000 = .5
County 2 = 2,400 ÷ 8,000 = .3
County 3 = 1,600 ÷ 8,000 = .2
Step 8.

Apply the ratio developed in Step 7 to the total unemployment estimate
for the LMA for the relevant time period. This will yield the estimate of
total unemployment in the county.

Example:
County 1 = 7,000 × .5 = 3,500
County 2 = 7,000 × .3 = 2,100
County 3 = 7,000 × .2 = 1,400

LAUS Program Manual 9-21

Use of 1990 Census Data in Disaggregating Labor Force Estimates—Census-Share Method

9-22 LAUS Program Manual

Annual

Processing

10 Annual Processing
I

n the current LAUS methodology, Handbook-based and model-based labor
force estimates are revised annually to take advantage of the latest available
information. New CPS population controls, revised Handbook components,
and revised State-supplied data are incorporated into the State and substate
estimates. In addition, State model performance is formally reviewed by both
State and national office staff, and adjustments are made to model specifications
when necessary. 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. The sections which follow discuss these processes in detail.

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 types of 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.)

LAUS Program Manual 10-1

Annual Model Review

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 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, SMD staff 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. SMD staff share 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 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.

10-2 LAUS Program Manual

Population Controls

Population Controls
CPS population controls are estimates of the working age civilian noninstitutional
population. These population estimates are derived by annual updates to the
resident U.S. population enumerated in the last decennial census through
components of population change based on a variety of administrative data.
Population controlling occurs when the sample-based monthly and annual average
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
model-based estimates, the monthly impact is via the CPS inputs to the model
estimates. 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 is accomplished through the additivity adjustment of substate estimates
to their respective statewide totals. Annual CPS population controls are
incorporated into CPS estimates through the revision of monthly and annual
average CPS labor force estimates at the end of the year. During annual
processing the revised CPS data are incorporated into the LAUS estimates in two
steps. First, the revised monthly CPS data are incorporated into the LAUS
monthly estimates when the models are re-estimated and smoothed. Second, the
revised State CPS annual average employment and unemployment levels are used
as control values to which the monthly model-based estimates are benchmarked.

How Population Estimates are Calculated
Current estimates of the national population for 16 categories of sex, race, and
Hispanic origin are derived by quarterly updates of the resident population
(enumerated in the last census) using components of population change. This
process is referred to as the “inflation-deflation” procedure, and uses the
following simple formula to update each category.
Revised Population

= Enumerated Base Population

+ Births to U.S. resident women
- Deaths of U.S. residents
+ Net international migration
+ Net movement of Armed Forces, Civilian Federal Employees, and their dependents to the U.S.

LAUS Program Manual 10-3

Population Controls

Births and deaths by sex, race, and Hispanic origin are obtained from the
National Center for Health Statistics (NCHS), generally through the
calendar year two years prior to the last July estimate date. Distribution
by sex, race and Hispanic origin is projected to the last July estimate date;
in the case of deaths, the projection is supported by data from the Current
Mortality Sample (10 percent of deaths). The projected distribution is
applied to a preliminary series of births and deaths, also obtained from
NCHS.
Estimates of international migration are made from several sources
including the Office of Refugee Resettlement, the Immigration and
Naturalization Service, and data on the net exchange of population with
the Commonwealth of Puerto Rico. The migration of Federally-affiliated
citizens is estimated using data provided by the Department of Defense
and the Office of Personnel Management.
The age distribution on the population estimate date is determined by
aging the population by quarter-year, subtracting by cohort, and adding
both types of net migration by cohort, for each age group, from the base
date forward. This procedure is known as the “cohort-component”
method for aging the population.
The population of the States must be estimated using less direct methods
than those used to derive population estimates for the Nation as a whole
because interstate migration, a large component of the change in State
populations, cannot be accounted for in as direct a way as births, deaths,
and legal immigration can. Two methods are used, and the official State
population estimates are an average of the two methods. State estimates
are “raked” proportionally to guarantee that they sum to the national total.
The first method is called the administrative records method because it
relies on the use of administrative records to derive estimates of the 0-64
and 65+ populations. Estimates of interstate migration for persons under
65 are derived from changes on IRS tax returns and the number of
exemptions claimed for persons under 65. Changes in the 65-and-older
population are derived from the change in the number of Medicare
enrollees.
The second method is called the components-of-change method because
it measures State population change as a composite of three components.
First is an estimate of migration of the school-age population based on
school enrollment data. Second is an estimate of the change in the 65and-over population based on changes in Medicare enrollments. Third is
a regression on changes in various symptomatic population indicators,

10-4 LAUS Program Manual

Population Controls

using a variety of data sources, including Federal income tax returns, automobile
registrations, and school enrollments, for estimating the change in the adult
population under age 65.
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, and persons under the age of 16.

State Annual Population Controls
Each January, the Census Bureau provides BLS with population estimates for
each State. These estimates are centered on the July “target” month for the
previous two years and annual averages for the previous three years. The most
recent July estimates are considered “provisional”, and estimates for the July prior
to that are considered “revised”. The July estimates two years before the
“provisional” estimates are also used to derive the intervening months, but should
not have changed since these estimates would have been finalized the year before.
The population estimates for the other months are derived using straight line
(linear) projections based on these three July estimates. The derivation of the
monthly values between the known (final, revised, and provisional) July estimates
is called “interpolation”. The derivation of the monthly estimates from the
provisional July estimate to the end of the most recent calendar year is called
“extrapolation”.
The following example illustrates the process. In January of 1994, BLS received
provisional July 1993 and revised July 1992 population estimates for each State.
The interpolation of these estimates may affect the LAUS estimates from August
1991 to July 1993. August 1991 is the starting point of the interpolation because
the July 1991 estimates would not change—those estimates were finalized the
year before. With interpolated estimates available for August 1991 through July
1993, estimates are still needed for August 1993 through December 1993. For
these months, the July 1991 estimates are extrapolated. This entire process results
in new statewide population estimates being available each January covering the
preceding two and one-half years. These new population estimates are then used
to adjust the State's CPS labor force data. This is done by multiplying each
month’s CPS employment and unemployment estimates 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 the
population controlling process because 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 CPS annual
averages are revised separately from the monthly estimates, as opposed to being
derived as the average of the re-controlled monthly estimates. Because separate

LAUS Program Manual 10-5

Population Controls

controls are applied to the monthly and annual data, the average of the
monthly CPS estimates will not necessarily equal the “official” CPS
annual average.

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. County total population
estimates are revised by the Census Bureau each year; place total
population estimates are revised every other year. The timing of the
availability of these data influences when these population estimates are
incorporated into the LAUS substate estimates.

10-6 LAUS Program Manual

Annual Re-Estimation or “Smoothing”

Annual Re-Estimation or “Smoothing”
At the end of each calendar year, the LAUS regression models are re-estimated,
smoothed, and benchmarked to the CPS annual average labor force estimates. Reestimation and smoothing is an integrated process which incorporates revised
State-supplied model inputs and revised CPS labor force data and involves a reestimation of all observations in a forward-back-forward manner. In this way,
every observation benefits from all of the data, past to present.
Each year States are provided the schedule for benchmarking activities in a
technical memorandum. States are instructed to provide the most up-to-date CES
nonagricultural wage and salary estimates reflecting the latest ES-202 benchmark.
They also provide revisions to UI continued claims and UCFE data, where
appropriate. Typically States submit these data during one of two benchmarking
cycles via the AP module in STARS. (See Chapters 11 and 12.)
In re-estimating and smoothing the models, the entire time series is used to reestimate every observation. The estimation process is run forward from the
beginning of the time series, then run backward so that earlier observations benefit
from later data, and then run forward again to the end of the year. It is possible to
perform this re-estimation and smoothing operation on the LAUS models because
they use a Kalman Filter to modify each of the model’s coefficients with the
addition of each monthly observation.
The Kalman Filter acts as filter, or weighting mechanism, which allocates how
much a model’s coefficients will change (and thus the estimates) with each new
period’s data. Because the Kalman Filter works by evaluating each successive
observation, one after another, the models can produce estimates in both a forward
and backward temporal direction. This means that the models can be “smoothed”,
a term used to describe the forward-back-forward re-estimation of the entire time
series. By producing estimates moving forward in time, each successive estimate
incorporates all of the information from the earlier months in the time series.
When the end of the time series is reached, 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
again, so that information from the first two passes can be incorporated into the
entire time series.
To the extent that the past and future help to describe the present, it is generally
easier and more accurate to fix a line to a period of data when you know the data
preceding and succeeding that period. In a similar fashion, it is easier and more
accurate to make an estimate for a month when you have estimates for the months
which preceded and succeeded that month. The forward-back-forward process in

LAUS Program Manual 10-7

Annual Re-Estimation or “Smoothing”

smoothing essentially provides this “pre” and “post” information for each
month, thus allowing a more accurate estimate to be made for every
month in the time series.

10-8 LAUS Program Manual

Benchmarking State Estimates

Benchmarking State Estimates
Benchmarking is a process of statistical adjustment conducted each year to adjust
the prior series of monthly LAUS model estimates to the CPS annual average
totals. This process controls the LAUS model-based estimates to independent
estimates with known levels of reliability—an 8 percent coefficient of variation on
the unemployment level, assuming an unemployment rate of 6 percent—which is
met for all States. The primary impetus for benchmarking the LAUS series to the
CPS is to give the estimates a measure of conformity appropriate to their use in
the distribution of Federal funds. Beyond this legislative impetus, benchmarking
LAUS estimates to the CPS is appropriate, given the role of the CPS in providing
a conceptual standard for the program.
The goal of LAUS benchmarking is twofold:
1. to

insure that the annual average of the final benchmarked series
equals the CPS annual average;

and
2. to

preserve the seasonal pattern of the model series as much as
possible.

This combination of goals places emphasis on selecting a statistical procedure that
minimizes the changes to the time series pattern. The Denton Method is used in
the LAUS benchmarking process because it is able to combine these two goals in
a single formula.
The Denton Method is a benchmarking methodology used to adjust estimates
from one series (monthly model-based LAUS) to another (CPS annual average
employment and unemployment). The monthly estimates are adjusted so that they
are exactly equal to the CPS annual estimates, while breaks between years and
distortions to the original month-to-month movements in the series are minimized.
This methodology uses a process based on constrained quadratic minimization.
This minimization routine is similar to the least squares idea of fitting a line to
data. Like least squares, which minimizes the sum of the squared differences
between the data points and a line fit to those points, the Denton procedure
minimizes the sum of squared percent differences between pairs of corresponding
months for the original and benchmarked estimates. The choice of this method is
particularly appropriate for the LAUS model estimates, since minimizing the
standard deviation of percent month-to-month changes was one of the original
selection criteria when the earliest models were developed. Maintaining the
month-to-month smoothness of the original estimates in the benchmarked series
was, therefore, an important consideration in choosing a benchmark methodology.

LAUS Program Manual 10-9

Benchmarking State Estimates

The Denton method accomplishes this minimization with a function
which penalizes differences between the original and benchmarked series
according to predetermined criteria. One of these criteria is to minimize
the sum of squared percent differences between corresponding month-tomonth changes in the original and benchmarked series. This portion of the
penalty function minimizes the changes to the seasonal pattern between
the original and benchmarked series. There is also a constraint which
requires that the average of the monthly estimates equals the CPS annual
average. To minimize breaks between years, successive groups of three
years are moved through the process, with one additional year added and
one dropped as the entire time series is benchmarked.
The Denton Method uses a mathematical formulation, called a
Lagrangian expression, to combine the penalty (constraint) functions
related to the benchmarked estimates into a set of benchmarking
equations. The Lagrangian expression incorporates a set of variables
called Lagrange multipliers, which, when the benchmarking equations are
partially differentiated, produces a set of weights for each month. The
weights times the CPS/model differences are summed and added to the
original model estimate to form the benchmarked estimate. While the set
of weights is based on the overall CPS/model difference for a given year,
the weights vary gradually from month to month as a result of solving the
equations to minimize the changes in the month-to-month differences
between the model and the benchmarked series.

Modification to the Denton Procedure: January 1998
The Denton approach provides a smooth link between December and
January of adjacent years with different overall benchmark adjustments.
Unfortunately, the original Denton method introduces a discontinuity
between the December benchmarked estimate of the last year and the
January estimate of the current year. To alleviate this problem, the method
was modified to provide a link to the current year model series. This was
done by adding an end point constraint (ENP) which forces the December
benchmark estimate to equal the model estimate, thereby preserving the
December-January model estimate of change.
Since the ENP constraints leaves the original series value for December of
last year unchanged, the adjustment to the annual benchmark for that year
must be spread over the 11 remaining months of that year and a smooth
link established for earlier years. To satisfy the annual constraint, the
own year weights must sum to one over the same 11-month period. These
adjustments make it somewhat more difficult to maintain the monthly
growth rate pattern in the original series for the first 11 months of the last
year.

10-10 LAUS Program Manual

Benchmarking State Estimates

While the ENP constraint could affect all the historical values of the adjusted
series, in practice the effects of this constraint converge to zero within 24 months
preceding the end point of the adjusted series.

LAUS Program Manual 10-11

Benchmarking Substate Area Estimates

Benchmarking Substate Area Estimates
As in regular monthly estimation, substate area benchmarking is limited
in data sources. However, substate estimates can be improved by
incorporating updated source data, revising prior inputs, adjusting for
changes in UI procedures and coverage, incorporating changes in
geographic definitions, and adjusting the areas to revised State estimates
through additivity. In some cases, such as Handbook agricultural
employment estimation, annual benchmarking of the series is most
critical, given the lack of ideal source data for generating monthly
changes at the labor market area level. Benchmarking of subState area
estimates is conducted during the first half of the each year for the
previous three year's estimates. Occasionally, more than three years are
revised, typically following changes in methodology or large revisions to
population data. The requirements and schedule for this activity are
provided to the States via LAUS technical memoranda. In addition,
specific instructions relating to the LSS software to be used in
benchmarking are provided to the regions and States, typically in January.

10-12 LAUS Program Manual

Benchmarking Substate Area Estimates

Incorporation of Substate Data Updates
The table below lists a number of Handbook inputs, the source of the updated
data, and the reason for the revision or update.

Input

Reason for update

Source

Handbook Employment:
Agricultural factors

new ALS data from USDA

BLS

CES employment

latest ES-202 benchmark

States

Step-3 ratios

latest CES benchmark

Atypicals

new data available

BLS
States

Handbook Unemployment:
Annual A & B factors

updated factors

BLS

UI Claims

revised UI counts

Survival rates

new age group populations

BLS

Entrant ratios 16-19

new age group populations

BLS

Entrant ratios 20 +

new age group populations

BLS

benchmarked to CPS

BLS

States

Additivity:
Statewide estimates

Disaggregation:
Employment/population ratios

new employment & population data

BLS

The two most important Handbook updates are the CES employment and the UI
claims data, both provided by the State.
The monthly sample-based employment estimates from the CES program
produces are benchmarked each year to more complete payroll counts from the
ES-202 program. CES estimates are revised back to the previous year’s
benchmark and brought forward, using link-relative extrapolation to the end of the
most recent calendar year. This means that employment estimates used in LAUS
estimation are typically revised for the preceding twenty-one months.

LAUS Program Manual 10-13

Benchmarking Substate Area Estimates

UI claimant data for substate areas are also revised at benchmarking time.
UI continued claims without earnings are reviewed by State staff for the
benchmarking period (usually the last three years). Area claims counts are
corrected, finalized, and checked against the statewide total. This is the
opportunity to correct errors and omissions uncovered during the course
of the year. It is also the time to incorporate the effect that changes in UI
law or practice have had on the LAUS estimates.
Revised Agricultural Labor Survey (ALS) annual change factors are
introduced at benchmarking. These factors are for the July-to-July period
ending in the most recently completed calendar year with new estimates
developed for the remaining months of the most recent year.
Consequently, the series revision for agricultural employment is for the
July two years ago to the most recent July, and extended for August to
December of the most recent year.
Revised Seasonal A and B Factors used to estimate Handbook
unemployed entrants, based on the previous years's CPS data, are issued
at benchmark time. The revised factors should be used, if possible, to
revise the previous year's estimates, as well as for monthly estimation in
the ensuing year.
Updated population survival rates, developed from actuarial tables from
the Department of Health and Human Services, are issued at
benchmarking to update population estimates used for area Youth
Population Ratios (YPRs). The updated ratios are not incorporated into
the estimates for the previous year, but instead are used only in the
upcoming year.
Employment-to-population ratios for employment disaggregation are
updated with revised population data from the Census Bureau P-25 and P26 series. County population data are revised each year and place
population data every other year.
In addition, benchmarking substate area estimates reflects benchmarking
of statewide LAUS estimates. The State series provides the control total
for additivity adjustment of area estimates.

Incorporating Changes in Geographical Areas
Changes in population levels, within metropolitan and nonmetropolitan
areas, and in cities and towns, can alter LAUS geographically defined
areas. Population growth in areas can result in the creation of new a labor
market area. Growth in areas contiguous to an LMA can result in a
redefined LMA. Cities and towns newly identified as having populations

10-14 LAUS Program Manual

Benchmarking Substate Area Estimates

of 25,000 or more are also included as LAUS estimating areas. Population
changes of this type are incorporated into the LAUS framework during the
benchmarking period.

Substate Processing
The BLS provides the States with title code listings (geographical listing of all
areas a state re-estimates), new ratios for disaggregation, a benchmark control file
(State revised estimates from STARS), new population data, and annual
agricultural factors, a and b factors, and survival rates.
States enter inputs, generate revised estimates, and transmit revised estimates to
the national office for review.
Revised substate data undergo a thorough review to insure accuracy and
consistency. The national office reviews and edits the State control and
disaggregation files—editing data relative to previous transmissions. Range
edits—increases or decreases of 20-25 percent and level shifts of 100—are
utilized to check the data against previous estimates, and consideration is also
given to a number of subjective factors, including the size of the area being
revised and current economic conditions.

LAUS Program Manual 10-15

Benchmarking Substate Area Estimates

10-16 LAUS Program Manual

STARS

Macro Output

11 STARS Macro Output
Introduction

A

ll states use the monthly macro called STARS (State Time Series Analysis
and Review System) each month to produce their employment and
unemployment rate estimates, and to transmit these to the BLS national

office.
In producing estimates, States provide input data such as CES employment,
strikers, and unemployment claims. These data are used not only to produce the
estimates, but are also stored in a national office database and are available
through an Extract Macro, along with data from the CPS, for use in various
analytical studies, model interpretation, and so forth.
Each time STARS is run, it provides both BLS and State analysts with output
containing a series of tables and graphs with information for studying
employment trends, preparing releases, or understanding the nature of a month-tomonth 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 analytical charts and tables.
• allow transmittal of the estimates to BLS.
CPS data are inputted at the national office, for access by the Macro, as soon as
the monthly national press release is issued. State analysts can enter their inputs
when they become available, review their listings, and redo them if an error is
found. States have the option to run estimates with preliminary numbers before
the actual data are available without transmitting the estimates to BLS, and, when
they are correct, final LAUS estimates are transmitted to BLS through the macro.

LAUS Program Manual 11-1

STARS Cover Sheet

STARS Cover Sheet
The cover sheet provides basic information
about the macro run. Starting at the top, it lets
the State analyst know if the estimates were
transmitted to BLS from the run, and if so,
when they were transmitted. It gives the State,
month, and year of the data for which the run
was made and shows the date and time of the
run. It lists the State-entered data items for the
run month and the prior month, allowing the analyst to quickly check
input data.
The State Comments section allows a State to document their run in the
output. Here an analyst can note the status of the input data or the type of
run he/she is making. For example, it can be noted whether a run is
preliminary or final. States are encouraged to use this section to identify
their runs. These comments are also very helpful to BLS analysts.
There is also a section for BLS notices. Included here are the current
production month, due dates, and special notes about the macro or data
transmittals. These messages are also displayed on the computer screen
when STARS is executed
Official Estimates Transmitted to BLS on 10/17/95 at 09:37:42 LAUS STARS Monthly
Estimation Tables (Version 1.0)
----------

Mississippi

----------

September 1995
Date: 10/12/95
Time: 15:09:06
-------------------------------------------------Data entered for September 95:
CES nonag w&s employment
CES adjustment for major strikes
State continued claims w/o earnings
UCFE without earnings

=
=
=
=

1,059,178
0
15,760
113

Data entered for August 95:
CES nonag w&s employment
=
1,051,708
CES adjustment for major strikes
=
0
State continued claims w/o earnings =
21,010
UCFE without earnings
=
115
--------------------------------------------------

11-2 LAUS Program Manual

STARS Cover Sheet

BLS Notices:
Estimates for the current production month (95/09) are
due on 10/17/95. If you need to transmit estimates for
a non-production month, contact your regional office.

LAUS Program Manual 11-3

STARS Table 1: Year-to-Date Model Estimates

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. This is
the place to look for the estimates created by a macro run. A quick
comparison of current estimates to earlier estimates can be made.
Developing trends and month-to-month changes can be observed. A
comparison of the seasonally adjusted and not seasonally adjusted series
can be made.

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 1: Year-to-Date Model Estimates
Mississippi, 1995

------------------------

Seasonally Adjusted
---- Employment ----

Date

Labor Force

-------------------- Unemployment --

Level

E/P

Level

Rate

JAN95

1,266,814

1,192,718

59.9

74,096

5.8

∫

∫

∫

∫

∫

∫

SEP95

1,269,626

1,183,546

58.9

86,080

6.8

------------------------

Not Seasonally Adjusted ---------------------- Employment ----

Date

Labor Force

-- Unemployment --

Level

E/P

Level

Rate

JAN95

1,254,822

1,177,416

59.1

77,406

6.2

∫

∫

∫

∫

∫

∫

SEP95

1,265,091

1,179,427

58.7

85,664

6.8

11-4 LAUS Program Manual

STARS Table 2: Changes from Prior Month/Year

STARS Table 2: Changes from Prior Month/Year
This table shows over-the-month and over-the-year changes for the each of the
four basic types of labor force estimates. Data for the month specified as
“current” in the macro session and the prior month are shown for 1978 through the
current year. The data for all years prior to the current year reflect the annual
updating. Both the level of change and the percent change are given.
This table is particularly useful for examining the normal seasonality of the
estimates. The current month-to-month change should be compared to changes
for the same months in previous years to see if the change is similar in magnitude
and in the same direction.
The over-the-year changes give an indication of the state's labor force trends, but
this can be better observed by looking at all months of historical data, available
from the extract macro or by viewing the plots of the monthly data provided in
Figures 1 & 2 of this monthly output.
Since year-to-year change does not contain seasonality, this change should be
about the same for both seasonally adjusted and unadjusted series.
LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 2a: Changes from Prior Month/Year
Mississippi Labor Force, September 1995
(Past year estimates are benchmarked)
------------------------

Seasonally Adjusted
--------------

------------------Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

1,012,531

1,016,151

-3,620

-0.4

.

.

∫

∫

∫

∫

∫

∫

∫

95

1,269,626

1,273,917

-4,291

-0.3

8,331

0.7

LAUS Program Manual 11-5

STARS Table 2: Changes from Prior Month/Year

------------------------

Not Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

1,011,397

1,012,881

-1,484

-0.1

.

.

∫

∫

∫

∫

∫

∫

∫

95

1,265,091

1,273,892

-8,801

-0.7

8,767

0.7

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 2b: Changes from Prior Month/Year
Mississippi Employment, September 1995
(Past year estimates are benchmarked)

------------------------

Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

944,081

943,715

366

0.0

.

.

∫

∫

∫

∫

∫

∫

95

1,183,546

1,190,882

-7,336

-0.6

2,706

11-6 LAUS Program Manual

∫
0.2

STARS Table 2: Changes from Prior Month/Year

------------------------

Not Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

938,325

937,392

933

0.1

.

.

∫

∫

∫

∫

∫

∫

95

1,179,427

1,187,928

-8,501

-0.7

2,189

∫
0.2

LAUS Program Manual 11-7

STARS Table 2: Changes from Prior Month/Year

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 2c: Changes from Prior Month/Year
Mississippi Unemployment, September 1995
(Past year estimates are benchmarked)

------------------------

Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

68,450

72,436

-3,986

-5.5

.

.

∫

∫

∫

∫

∫

∫

∫

95

86,080

83,035

3,045

3.7

5,625

7.0

------------------------

Not Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

73,072

75,489

-2,417

-3.2

.

.

∫

∫

∫

∫

∫

∫

∫

95

85,664

85,964

-300

-0.3

6,578

8.3

11-8 LAUS Program Manual

STARS Table 2: Changes from Prior Month/Year

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 2d: Changes from Prior Month/Year
Mississippi Unemployment Rate, September 1995
(Past year estimates are benchmarked)

------------------------

Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

6.8

7.1

-0.3

-4.2

.

.

∫

∫

∫

∫

∫

∫

∫

95

6.8

6.5

0.3

4.6

0.4

6.2

------------------------

Not Seasonally Adjusted
--------------

------------------

Change

-------------

From Prior Month

From Prior Year

----------------

---------------

Year

September

August

Number

Percent

Number

Percent

78

7.2

7.5

-0.3

-4.0

.

.

∫

∫

∫

∫

∫

∫

∫

95

6.8

6.7

0.1

1.5

0.5

7.9

LAUS Program Manual 11-9

Table 3: Components of Change

Table 3: Components of Change
The components of change tables show the year-to-date current model
estimates and the change in these estimates from the prior month. It also
breaks the change down, showing the over-the-month change in each
model variable component. This is done for each of the models-employment-to-population ratio in Table 3a and unemployment rate in
Table 3b. The components of change are calculated using the same
coefficients for both months involved in the over-the-month change. The
current month's coefficient for each variable is used.
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 their past
behavior. A complete listing of historical components of change is
available through the extract macro.

11-10 LAUS Program Manual

Table 3: Components of Change

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 3a: Components of Change
Employment-to-Population Ratio
Mississippi, 1995

Current

Chg from

-Regression-

-----Residual----

Date

Month

Prior Mo

CESEP

Trend

Seasonal

JAN95

59.12

-1.07

-0.73

0.37

-0.72

∫

∫

∫

∫

∫

∫

SEP95

58.74

-0.51

0.19

-0.36

-0.34

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 3b: Components of Change
Unemployment Rate
Mississippi, 1995

Current

Chg from

-Regression-

-----Residual----

Date

Month

Prior Mo

Claims Rt

Trend

Seasonal

JAN95

6.17

0.79

0.39

-0.06

0.42

∫

∫

∫

∫

∫

∫

SEP95

6.77

0.02

-0.37

0.06

0.33

LAUS Program Manual 11-11

Table 4: Components of the Signal

Table 4: Components of the Signal
Table 4 provides current year values for the regression explanatory
variable, its coefficients, and the time series components: residual trend
and residual seasonal. This table shows what portion of the model
estimate is composed of each of its component parts.
These data help analysts to interpret the components of change table by
providing the data necessary to sort out what change in the explanatory
variable is due to the changing coefficient and what change is due to the
change in the input variable value.
LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 4a: Components of the Signal
Employment-to-Population Ratio
Mississippi, 1995
------Regression------

-----Residual-----

--Coeffs--

--Inputs--

Date

CESEP

CESEP

Trend

Seasonal

JAN95

0.5512

52.6792

30.3939

-0.3138

∫

∫

∫

SEP95

0.5721

52.7482

∫
29.0162

∫
-0.4556

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 4b: Components of the Signal
Unemployment Rate
Mississippi, 1995
------Regression------

-----Residual-----

--Coeffs--

--Inputs--

Date

Claims Rt

Claims Rt

Trend

Seasonal

JAN95

1.0811

2.2109

4.1792

-0.4007

∫

∫

∫

SEP95

0.6908

1.4986

11-12 LAUS Program Manual

∫
5.0718

∫
0.6643

Table 5: Input Data

Table 5: Input Data
Table 5a provides current year CPS data relevant to the models estimates. Table 5b
provides current year model input data provided by the State. These data are
useful in analyzing model estimates.
LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 5a: CPS Data
Mississippi, 1995

OBS

DATE

CPSRT

CPSEP

CPSUN

CPSEM

CPSPOP

1

DEC94

6.04

59.91

76600

1192300

1990000

∫

∫

∫

∫

∫

∫

∫

10

SEP95

7.15

59.46

91900

1194000

2007989

LAUS STARS Monthly Estimation Tables (Version 1.0)
Table 5b: State Data
Mississippi, 1995

OBS

DATE

CNTWOER

UCFE

UIATYP

HBEXH

CESEM

STRIKERS

1

DEC94

15934

189

0

.

1068857

0

∫

∫

∫

∫

∫

∫

∫

∫

10

SEP95

15760

113

0

.

1059178

0

LAUS Program Manual 11-13

Figures 1 and 2: Plots of Employment Level and Unemployment Rate

Figures 1 and 2: Plots of Employment Level and
Unemployment Rate
Figures 1 and 2 plot three years of historical data and current-year
estimates up through the latest month of complete input data. Each figure
displays two plots: the top plot shows the seasonally adjusted model
estimate, and the bottom plot shows the not seasonally adjusted model
estimate and the CPS.
The plots allow analysts to see how closely the model is following the
CPS for a particular month and over time. It also allows the analyst to see
seasonal patterns in the data and observe the trend/cycle movements of
estimates in recent years.

11-14 LAUS Program Manual

Figures 1 and 2: Plots of Employment Level and Unemployment Rate

LAUS STARS Monthly Estimation Tables (Version 1.0)
Figure 1:

Mississippi Employment

(Past year estimates are benchmarked)
Model Employment (in thousands):
CPS
Employment (in thousands):

S
E
A
S
A
D
J

*
o

|
|
|
|
|
1,220 +
|
|
|
|
|
|
|
|
|
|
|
|
**
*
|
1,182 +
|
|
**** **| **
* *
|
|
|
| * **
|
**
|
|
|
*****
|
|
1,144 +
|
***
|
|
|
|
| ** ***
|
|
|
|
**
|
|
|
1,106 +
* **|
|
|
|
|
****
|
|
|
|
| **** *
|
|
|
|
1,068 +
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1,030 +
|
|
|
|
---+-------------+-------------+-------------+-------------+-JAN92
JAN93
JAN94
JAN95
JAN96
Date

N
O
T
S
E
A
S
A
D
J

|
|
|
|
|
1,220 +
|
|
o
|
|
|
|
|
o oo
|
oo
|
|
|
|
o*
*|
* o
|
1,182 +
|
|
* *** * ***
**
|
|
|
oo o
*|
*
o oo * **
|
|
|
o * * |***
| oo
|
1,144 +
|
o****
o* o
|
|
|
o |
*
|
|
o
|
|
o o **|***
o o
|
|
1,106 +
o*o * o*ooo
|o
|
|
|
* **
|
|
|
|
|
*** *
o
|
|
|
1,068 + *ooo o
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1,030 + o
|
|
|
|
---+-------------+-------------+-------------+-------------+-JAN92
JAN93
JAN94
JAN95
JAN96
Date

NOTE: 6 obs hidden.
When the model and the CPS have similar values, only the model
symbol (*) is shown. The CPS symbol is "hidden" behind it.

LAUS Program Manual 11-15

Figures 1 and 2: Plots of Employment Level and Unemployment Rate

LAUS STARS Monthly Estimation Tables (Version 1.0)
Figure 2:

Mississippi Unemployment Rate

(Past year estimates are benchmarked)
Model Unemployment Rate:
CPS
Unemployment Rate:

S
E
A
S
A
D
J

N
O
T
S
E
A
S
A
D
J

*
o

|
|
|
|
|
11.1 +
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9.7 +
|
|
|
|
|
|
|
|
|
|
*
|
|
|
|
8.3 + **** * ***
|
|
|
|
|
*
|
|
|
|
|
* |
|
|
|
6.9 +
*|
|
* **|
*
|
|
**** *
** ****** *****
|
*
|
|
|
***
|
*
**
|
5.5 +
|
|
| *
*
|
|
|
|
|**
|
|
|
|
|
|
4.1 +
|
|
|
|
---+-------------+-------------+-------------+-------------+-JAN92
JAN93
JAN94
JAN95
JAN96
Date
|
|
|
|
|
11.1 +
o
|
|
|
|
|
|
|
|
|
|
*
|
|
|
|
9.7 +
o
|
|
|
|
|
o
|
|
|
|
| ***
**
|
|
|
|
8.3 +
*
*
|
|
|
o
|
|
o *
oo
| o
*
|
*
|
o
|
|
oo
|
o
oooo oo
o |
*
|
6.9 +
*
** o
***
**
|
o **
|
|
* | ** * ***
| * *
* * *
|
|
*|
|
* *o
* *
|
5.5 +
o o
o o* **|
o o
|**o
|
|
|
|
|oo*
|
|
|
o |
|
|
4.1 +
|
|
o
|
|
---+-------------+-------------+-------------+-------------+-JAN92
JAN93
JAN94
JAN95
JAN96
Date

NOTE: 12 obs hidden.
When the model and the CPS have similar values, only the model
symbol (*) is shown. The CPS symbol is "hidden" behind it.

11-16 LAUS Program Manual

STARS

User’s Guide

12 STARS User’s Guide at SunGard
Introduction

T

he computer interface used by the States to make, review, and transmit their
model estimates is a system of interactive programs called STARS (State
Time Series Analysis and Review System).

To access STARS, States log on to the mainframe computer at the SunGard
Computing Services in Voorhees, New Jersey. All STARS-related sessions begin
with logon procedures that are described below. (Analysts should use only the
account/initials specified by BLS, since only those user IDs will permit access to
the STARS system).

Instructions for Logging on SunGard to Access STARS
Communication Settings:

Phone Number: (800) 811-5861 or (301) 853-9200
Terminal Emulation: VT-100
Baud Rate: 300, 1200, 2400, or 9600
Data Bits: 7
Stop Bits:1
Parity: Even
Duplex (Local Echo): Half (On)
Logon Steps:
1.) After a connection is made, press the key once. The following
prompt will appear on your screen:

LAUS Program Manual 12-1

Introduction

WELCOME TO THE SUNGARD
COMPUTER SERVICES NETWORK
ENTER APPLICATION NAME ===>
2.) Enter in SPWYL and press the key.
Enter your SunGard USERID, account, and password at the appropriate
prompts.

To execute STARS, type:
CALL FROM &&&YBDBLS.A130.@@STARS
The first screen identifies the State associated with the user's ID and the
latest period for which CPS data are available during this STARS session.
It also displays information about LAUS production months and due
dates. An example is shown below.

Laus State Time Series Analysis and Review System (STARS)
Welcome to STARS, Version 1.0
State: State Name
CPS data available through 95/06.
The current LAUS production months are:
Preliminary: 95/06
Revised:

95/05

Statewide estimates for these months can be transmitted
beginning 07/07/95. They are due on 07/18/95. If you need
to transmit estimates for months other than these, contact
your regional office. For a full set of official due
dates, see the monthly LAUS schedule.
Press Enter to continue...

12-2 LAUS Program Manual

Main Options

Main Options
The next screen displays the main options available in STARS: 1) create new
estimates, 2) review current estimates, 3) transmit current estimates, and 4) invoke
the annual processing module. These options, which are described in detail below,
appear as shown in the example below. Note that users are asked to verify their
choice (so as to prevent, for example, accidental transmission). The default for the
verification prompt can be chosen simply by pressing the Enter key. In the
example below, the default response is “Y” for yes.

------------Options-----------1. Create new estimates
2. Review/Extract estimates
3. Transmit current estimates
4. Annual Processing
5. Exit STARS

Enter choice (#): 1

Are you sure you want to create new estimates? (Y/n):

NOTE: In all STARS software, whenever there is a “yes/no” type choice at a
prompt (as in the verification prompt above), the default response is shown with
an uppercase letter (Y in the example above). However, the actual case of the letter
entered by the user in response is not important.

LAUS Program Manual 12-3

STARS Estimation Module

STARS Estimation Module
Creating New Estimates
Option 1 of the STARS main options menu allows the user to create new
estimates. When choosing this option, the user must specify the month for
which the estimates are to be created. The default for this prompt is the
current LAUS preliminary production month. Note that estimates can be
created for the current year only (revised inputs can be entered for
December of the prior year to improve the estimates for the current year) .
Also note that estimates are always made through the last month for
which complete input data exist. When the specified month differs from
the last month for which data are available, it is the selected month which
is used to determine which data are displayed on certain STARS output
tables. (See the Tables section below.)

Specify the year and month (numeric) for which you want
to create new estimates. The default is 95/06

YY/MM:

After choosing the estimation month, the user can then choose to update
the model input data. The primary State input data are CES
nonagricultural wage & salary employment (CESEM) and the State UI
continued claims count without earnings (CNTWOER). The secondary
State input data are adjustments for major strikes and the UCFE claims
count without earnings. Updates can be made for zero, one, or two
months. The default (which can be chosen simply by pressing the Enter
key) is to update data for two months--the specified month and the month
before. As shown in the example below, STARS displays specifically
which month(s) the user may choose to update.

---------

Data Updates

--------

Enter updates for ...
1. June 95 and May 95

12-4 LAUS Program Manual

(default)

STARS Estimation Module

2. June 95
3. No updates, use current inputs

Enter choice (#):

Entering Data
When entering data updates, the user will be prompted for the input data for that
State. The default response for these prompts is that whatever data are currently in
the State database will be used (i.e., “no change”). An example of the data input
screen follows.

Notice:
States are asked to double check their CNTWOER and
UCFE claims counts to be sure that commuter and
interstate claims have been include in these data
------ Enter Input Data -----Jun 95 CES nonag w&s employment..............2426095
May 95 CES nonag w&s employment..............2403023
Jun 95 CES adjustment for major strikes......258
May 95 CES adjustment for major strikes......
Jun 95 UI continued claims w/o earnings......31525
May 95 UI continued claims w/o earnings......33155
Jun 95 UCFE claims w/o earnings..............
May 95 UCFE claims w/o earnings..............

When all of the input data have been entered, STARS displays them for review.
Note two things in the example below. First, if “no change” is selected for any
observation, STARS issues a warning message that estimates will not be made for
months in which any of the input data are missing. In other words, to make an
estimate, there must be a complete series of input data. Second, an entry can be
changed by selecting the corresponding line number; STARS issues a prompt for
the user to re-enter the observation

LAUS Program Manual 12-5

STARS Estimation Module

.
You have entered the following input data or State Name:
Variable

Year

Month

Data

--

#

-------------------------------

----

-----

--------

1.

CES nonag w&s employment

95

06

2426095

2.

CES nonag w&s employment

95

05

2403023

3.

CES adjustment for major strikes

95

06

258

4.

CES adjustment for major strikes

95

05

NO CHANGE

5.

UI continued claims w/o earnings

95

06

31525

6.

UI continued claims w/o earnings

95

05

33155

7.

UCFE claims w/o earnings

95

06

NO CHANGE

8.

UCFE claims w/o earnings

95

05

NO CHANGE

Note: You have entered NO CHANGE for at least one observation. If
that observation is currently missing, STARS will not make an estimate for that or any succeeding month.

Enter the line number of an entry you want to change, or press the
ENter key to continue:
7

95/06 UCFE claims w/o earnings................ 73

STARS displays all of the data for review after each change, until no
changes are made.
In some cases, a user might want to make an estimate without saving
either the input data or the estimate in the State database (such as, when
making a projection for the current month based on the current CPS and
the previous month's State-supplied input data). To accommodate this,
STARS allows the user to choose whether or not the database should be
updated. An example of the prompt appears below. However, on the
assumption that most estimates will be using official input data, the
default (chosen by pressing the Enter key) is “Yes.”
Do you want to update your database with the inputs you
have just entered and the resulting model estimates?
(Y/n):

12-6 LAUS Program Manual

STARS Estimation Module

Estimation
Before STARS begins estimation, the user is given the opportunity to record notes
or comments. These will appear on the STARS output tables. (See the Tables
section below.) An example appears below.
Enter your notes or comments below, up to a maximum of 5
lines and 50 characters per line (even though a line is
approximately 80 characters long). Press the Enter key at a
new line prompt when you are finished.
Line 1: This is a test of the new STARS
Line 2: estimation at SunGard
Line 3:

Please wait. The job is running now.
Approximate run time is 1 to 3 minutes.
Please do not PRESS any key OR program will be interrupted

WAIT started at 8:58:45 a.m.
WAIT ended at 8:59:14 a.m.

After notes/comments (if any) have been entered, STARS automatically submits a
job to the central processor for estimation. While the job is being executed, the
system enters a wait state as indicated by the message “WAIT started at... .”
During the wait state, the user should not touch the keyboard. When the job is
complete, the output tables are fetched into the user's active workspace, where
they can be viewed or printed. All STARS runs are saved under the user’s account/
initials for later access. As shown in the example below, the name of the output
file is displayed, and the user can choose to list the tables simply by pressing the
Enter key.
The job is complete.

The output tables are in your active workspace and
have been saved as STARS.JUN95.D950718.T08813.

Do you want to list the tables on your screen?
(Y/n):

LAUS Program Manual 12-7

STARS Estimation Module

Tables
STARS monthly tables provide information about the model estimates.
The tables are:
Cover Sheet
Table 1:

Year-to-Date Model Estimates

Table 2:

Changes from Prior Month/Year
a. Labor Force
b. Employment
c. Unemployment
d. Unemployment Rate

Table 3:

Components of Change
a. Employment-to-Population Ratio
b. Unemployment Rate

Table 4:

Components of Signal
a. Employment-to-Population Ratio
b. Unemployment Rate

Table 5:

Input Data
a. CPS Data
b. State Model Input Data

Figure 1:

Plots of Employment Level

Figure 2:

Plots of Unemployment Rate

Tables 1, 3, 4, and 5 display current-year estimates up through the latest
month of complete input data. Table 2 shows the over-the-month and
over-the-year change, in both the current year and all previous years, for
the month specified by the user, which may differ from the latest month of
available estimates. (See the Creating New Estimates section above.)
Figures 1 and 2 plot three years of historical data and current-year
estimates up through the latest month of complete input data, for the
employment level and unemployment rate, respectively. Each figure
contain two plots: the top plot shows the seasonally-adjusted estimates
(model E/P converted to employment level and model unemployment

12-8 LAUS Program Manual

STARS Estimation Module

rate, respectively); the bottom plot shows the not seasonally-adjusted estimates,
along with the CPS. Unless otherwise noted, all model-related data are forwardfilter estimates, i.e., the data are not smoothed or benchmarked.

LAUS Program Manual 12-9

STARS Review Module

STARS Review Module
Review Current Estimates—Extract Data
Option 2 of the STARS main options menu allows the user to review
current estimates and extract both current and historical data. The first
data review option is to run the Data Extract module, which is an
interactive program that lets users access model and CPS data in their
State databases.
The module begins by showing the State for which data are going to be
extracted. An example using the State of Connecticut (CT) is shown on
the following pages.
------- Review/Extract Options ------1.

Extract data (default)

2.

List latest STARS estimation run

3.

Exit review

Enter choice (#): 1

Specifying Years
The user is then prompted to enter the beginning and ending years for
which data are desired. Data are available monthly from January 1978
through the month for which the most recent CPS data are available. The
default beginning year is 78 and can be chosen simply by pressing the
Enter key (in response to that prompt). The default ending year is the
current year. If a question mark is entered in response to either prompt,
the valid years are displayed, as in the example below.
Enter beginning (2-digit) year

?

Valid years: 78 - 95
Enter beginning (2-digit) year [78]: 89
Enter ending (2-digit) year [95]:

12-10 LAUS Program Manual

92

STARS Review Module

Specifying Variables
The user must then enter the desired variables. The “variables” are simply the
names of individual data series, e.g., “CPSEM” refers to the series of monthly
CPS employment observations. The prompt is:
Enter variable list below (or “/” for help).
- There is a maximum of 8 variables/extract.

Variable (s): ?

Commas or blanks must separate variables in a list; up to eight variables can be
extracted in one run. If a question mark is entered in response to this prompt,
instructions and a list of valid variables are displayed as shown below.
Enter variable list below (or “?” for help).
Variable (s): ?
Enter the variable or variables (using blanks or commas between variables in a list). Valid variables are listed below and are described in
detail in the file &&&BDBLS.A130.STARS.VAR.DES.

BCESEP

BCLRADJ

BCLRFADJ

BCLRST

BCLSTFE

BHBEXHRT

CESADJ

CESEM

CESEP

CLRADJ

CLRFADJ

CLRST

CLRSTFE

CNTWOER

CPSEM

CPSEP

CPSLF

CPSPOP

CPSRT

CPSUN

DENTRT

EMLAUS

EMLAUSSA

EMSAF

EMSIGSMT

EMSIGUPD

EPLAUS

EPLAUSSA

EPREGST

EPREGUPD

EPSEASMT

EPSEAUPD

EPSIGSMT

EPSIGUPD

HBEXH

HBEXHRT

LFSIGUPD

NENTRT

NEWEN

REEN

RENTRT

SENTRT

STPH

STSE

STSEUF

STSEUFPH

STTA

STUF

STRIKERS

UCFE

UIATYP

UNLAUS

UNLAUSSA

UNSAF

UNSIGSMT

UNSIGUPD

URLAUS

URLAUSSA

URREGSMT

URREGUPD

URSEASMT

URSEAUPD

URSIGSMT

Group:

LAUS

LFLAUS

EMLAUS

EPLAUS

UNLAUS

URLAUS

LAUSSA =

LFLAUSSA

EMLAUSSA

EPLAUSA

UNLAUSA

URLAUSSA

SMT

=

LFSIGSMT

EMSIGSMT

EPSIGSMT

UNSIGSMT

URSIGSMT

CPS

=

CPSLF

CPSEM

CPSUN

CPSRT

EM

=

CPSEM

EMLAUS

CESEM

=

NOTES: 1. Groups cannot be selected with other variables or groups
2. Some variables are State-specific

LAUS Program Manual 12-11

STARS Review Module

In addition to the file noted above, variable descriptions also appear on
pages 12-16 through 12-18 of this Chapter.
An example of entering variables is shown below. Note that each entry is
checked for validity; if it is not valid, the response must be corrected.
Enter variable list below (or “?” for help).
- There is a maximum of 8 variables/extract.

Variable(s): CLRST URLAUS NENRT

’NENTRT” is not valid.
Enter new variable: NENTRT

Reviewing Selections
After the years and variables have been entered, the choices are displayed
for review and confirmation. The default response (which can be chosen
just by pressing the Enter key) is to extract the specified data.
*

STARS Data Extract Module

*

State............Connecticut
Beginning Year...89
Ending Year......92
Options:

1. Extract specified data (default)
2. Return to the variable list
3. Start over
4. Exit extract

Enter choice (#):

Data Extract Output
If the “extract specified data” option is chosen, STARS automatically
submits a job to the central processor to read the desired data from the
State database. When the job is complete, the formatted output tables are
fetched into the user's active workspace, where they can be viewed or
printed. The tables are also saved on temporary storage. In addition, the
data are saved in a text file for downloading. (See the Text File section
below.)

12-12 LAUS Program Manual

STARS Review Module

As shown in the example below, the names of these files are displayed, and the
user can choose to list the output tables simply by pressing the Enter key. (Note
that the alphabetical FIPS code that appears in the file name in this example is
CT.)
The

job is complete.

The output tables are in your active workspace and have
been saved as EXTRACT.CT.D950718.T094454 on WORK16.

A text file of the extracted data (for downloading) has
been saved as EXTRACT.CT.DATA.D950718.T094454 on WORK16.

Do you want to list the tables on tour screen?
(Y/n):

For the job shown in the examples above, the output is:
LAUS STARS Data Extract Tables
--------

Connecticut

--------

YEAR

MONTH

CLRST

URLAUS

NENTRT

89

1

1.7

3.6

2.1

89

2

1.9

3.6

2.1

89

3

1.8

3.1

1.9

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

92

10

2.3

6.5

2.4

92

11

2.4

6.3

2.4

92

12

2.5

6.1

2.3

At the end of each run, the user is given the opportunity to run the module again.
The default response, no, will cause the module to exit to WYLBUR. Note that
each run saves the output tables and data with a time and date stamp, thus creating
unique files.
Do you want to run the Data Extract Module again?
(y/N):

LAUS Program Manual 12-13

STARS Review Module

Text File
Each run of the Data Extract module saves the extracted data in a text file,
so that it can be downloaded or read into another program. The first two
columns in this file are for the 2-digit year and month. The data follow in
columns 12 digits wide, with each column separated by a comma and
space. Note that observations are padded on the left with zeroes, so that
each column maintains a constant width. Thus, the file layout is:
Variable

Columns

Year

1-2

Month

5-6

Variable #1

9-20

Variable #2

23-34

Variable #3

37-48

Variable #4

51-62

Variable #5

65-76

Variable #6

79-90

Variable #7

93-104

Variable #8

107-118

For the job shown in the examples above, a line from the text file (with
column numbers shown for reference) is:
1

2

3

4

5

6

7

1234567890123456789012345678901234567890123456789012345678901234567890
89, 01, 0000000001.7, 0000000003.6, 0000000002.1

After downloading, the data can be imported with the following
commands:
LOTUS 1-2-3 FOR DOS

MICROSOFT EXCEL (WINDOWS)

1. /File

1. File

2. Import

2. Open

3. Numbers

3. {Enter filename}

4. {Enter filename}

4. Text (“Text File Options” box)

5. {Enter}

5. Comma (column delimiter)
6. OK (twice)

12-14 LAUS Program Manual

STARS Review Module

Missing values in the text file are indicated by a single period (“.”). Lotus ignores
missing values, shifting data columns to the left; Excel imports missing values as
a character string (label).

LAUS Program Manual 12-15

STARS Review Module

STARS Variable Descriptions
Variable

Description

BCESEP

Model coefficients for CESEP

BCLRST

Model coefficients for CLRST

BCLRSTFE

Model coefficients for CLRSTFE

BHBEXHRT

Model coefficients for LAUS Handbook exhaustee rate

CESADJ

CESEM adjusted for strikers

CESEM

CES State total nonagricultural wage and salary employment

CESEP

CES employment-to-population ratio, = 100*CESADJ/CPSPOP

CLRST

Claims rate, adjusted for strikers, = 100*CNTWOER/CESADJ

CLRSTFE

Claims rate, including UCFE, adjusted for strikers,
= 100*(CNTWOER + UCFE)/CESADJ

CNTWOER

Continued claims for unemployment insurance without earnings

CPSEM

CPS State employment

CPSEP

CPS employment-to-population ratio, = 100*CPSEM/CPSPOP

CPSLF

CPS State labor force

CPSPOP

CPS State population (on latest population control)

CPSRT

CPS State unemployment rate

CPSUN

CPS State unemployment level

DENTRT

CPS census division entrant rate,
= 100*(CPS entrants/(CPS entrants + CPS employment))

EMLAUS

Model employment; “spliced” EMBMK plus current year-to-date EMSIGUPD

EMLAUSSA

Model employment, seasonally adjusted; = EMLAUS / EMSAF (multiplicative factor) or = EMLAUS - EMSAF (additive factor)

EMSAF

Employment seasonal adjustment factor; can be multiplicative or additive

EMSIGSMT

Historical model employment, smoothed but not benchmarked

EMSIGUPD

Forward-filter (not smoothed) model employment,
= EPSIGUPD*(CPSPOP/100)

EPLAUS

Model employment-to-population ratio, = 100*EMLAUS/CPSPOP

EPLAUSSA

Model employment-to-population ratio, seasonally adjusted,
= 100*EMLAUSSA/CPSPOP

EPREGSMT

Employment-to-population ratio model smoothed regression component

EPREGUPD

Employment-to-population ratio model forward-filter regression component

EPSEASMT

Employment-to-population ratio model smoothed seasonal component

12-16 LAUS Program Manual

STARS Review Module

STARS Variables Descriptions (Continued)
Variable

Description

EPSEAUPD

Employment-to-population ratio model forward-filter seasonal component

EPSIGSMT

Employment-to-population ratio model smoothed signal

EPSIGUPD

Employment-to-population ratio model forward-filter signal

EPTRDSMT

Employment-to-population ratio model smoothed trend component

EPTRDUPD

Employment-to-population ratio model forward-filter trend component

HBEXH

LAUS Handbook exhaustees

HBEXHRT

LAUS Handbook exhaustee rate, = 100*HBEXH/CESEM

JOBLEV

CPS job leavers

JOBLOS

CPS job losers

LFLAUS

Model labor force; “spliced” LFBMK plus current year-to-date LFSIGUPD

LFLAUSSA

Model labor force, seasonally adjusted

LFSIGSMT

Historical model labor force, smoothed but not benchmarked

LFSIGUPD

Forward-filter (not smoothed) model labor force

NENTRT

CPS national entrant rate (see DENTRT for equation)

NEWEN

CPS State new entrants

REEN

CPS State re-entrants

RENTRT

CPS census region entrant rate (see DENTRT for equation)

SENTRT

CPS State entrant rate (see DENTRT for equation)

STFIPS

Two-digit State FIPS code

STPH

CPS State nonagricultural private household employment

STRIKERS

State estimates of number of persons who did not work during the pay period
including the 12th of the month because their union was on strike or because
they were unwilling to cross a union picket line

STSE

CPS State nonagricultural self-employed employment

STSEUF

CPS State nonagricultural self-employed and unpaid family employment

STSEUFPH

CPS State nonagricultural self-employed, unpaid family, and private household employment

STTA

CPS State total agricultural employment

STUF

CPS State unpaid family employment

UCFE

Continued claims without earnings under the unemployment compensation
for federal employees program

UIATYP

Atypical UI claims adjustment

UNLAUS

Model unemployment; “spliced” UNBMK plus current year-to-date UNSIGUPD

LAUS Program Manual 12-17

STARS Review Module

STARS Variables Descriptions (Continued)
Variable

Description

UNLAUSSA

Model unemployment, seasonally adjusted; = UNLAUS / UNSAF (multiplicative factor) or = UNLAUS - UNSAF (additive factor)

UNSAF

Unemployment seasonal adjustment factor; can be multiplicative or additive

UNSIGSMT

Historical model unemployment, smoothed but not benchmarked

UNSIGUPD

Forward-filter (not smoothed) model unemployment, = EMSIGUPD / [(100/
URSIGUPD) - 1]

URLAUS

Model unemployment rate; “spliced” URBMK plus current year-to-date URSIGUPD

URLAUSSA

Model unemployment rate, seasonally adjusted

URREGSMT

Unemployment rate model smoothed regression component

URREGUPD

Unemployment rate model forward-filter regression component

URSEASMT

Unemployment rate model smoothed seasonal component

URSEAUPD

Unemployment rate model forward-filter seasonal component

URSIGSMT

Unemployment rate model smoothed signal

URSIGUPD

Unemployment rate model forward-filter signal

URTRDSMT

Unemployment rate model smoothed trend component

URTRDUPD

Unemployment rate model forward-filter trend component

Review Current Estimates -- Latest STARS
Estimation Run
Note: This option (Option 2 of the STARS main option menu) is not
currently available at SunGard.

12-18 LAUS Program Manual

STARS Transmit Module

STARS Transmit Module
Transmit Current Estimates
Option 3 of the STARS main options menu allows the user to transmit current
estimates, which officially sends the estimates to BLS files. This option should be
used only after the preliminary and revised estimates for the two production
months shown have been reviewed. States should never transmit unreviewed
estimates, for the estimates transmitted are considered official.
The data and tables transmitted to BLS contain preliminary estimates for the
current month and revised estimates for the previous month. The two months
involved are determined according to the Monthly LAUS Schedule. In addition to
transmitting the current and previous month's data, the macro will transmit
corrections made to data for earlier months (of the current year). Corrections to
data for earlier months may be made only after contacting the regional office since
special arrangements must be made to allow data for non-production months to be
read to a State's database.
If the transmit option is chosen, the user will be asked to verify the choice. The
default (chosen simply by pressing the Enter key) is “Yes.” An example of this is
shown below.
*** Transmitting officially sends your estimates ***
*** to BLS files. Please be sure that your last

***

*** estimation run is for the current production ***
*** month (95/06).

***

ARE YOU SURE YOU WANT TO TRANSMIT? (Y/n):

If the user does indeed wish to transmit, the State and time period for which
estimates will be transmitted are displayed. To transmit these estimates, the user
must again press the Enter key. This double check on transmission was added to
reduce the number of accidental or erroneous transmissions

LAUS Program Manual 12-19

STARS Transmit Module

.
*

STARS Transmit Module

*

State................Oklahoma
Transmitting for.....JUN95

Press Enter to continue ...

If the month specified in the previous estimation run was not the current
production month for preliminary estimates, or if preliminary estimates
have already been transmitted for the production month, the Transmit
module will not allow transmission unless special approval has been
granted.
If preliminary estimates for the production month are being transmitted
for the first time, or if special approval has been granted, the Transmit
module submits a job to transfer the estimates to BLS files for processing.
Also, the STARS output tables are annotated as having been officially
transmitted, and the date and time of the transmission are added to the
listing. The transmission is noted in a transmit log, and a copy is produced
for BLS review.

12-20 LAUS Program Manual

STARS Annual Processing Module

STARS Annual Processing Module
Annual Processing
At the end of each calendar year, the BLS performs a series of activities known
collectively as Annual Processing (AP). Annual processing includes re-estimation
(smoothing) of the models, benchmarking to CPS annual averages, and reseasonal adjustment. The first step in the process is to incorporate revisions to
State-supplied model input data. Option 4 of the STARS main menu runs the
Annual Processing (AP) module, which is an interactive program that lets users
enter revisions to their State input data. This option is available only during the
annual processing period.
The first screen of the AP module displays the AP options. An example is shown
below.
*

STARS Annual Processing

*

------------- Annual Processing options ------------1. List........Create a review listing
2. Correct.....Modify a month/year value for a series
and create a correction file
3. Transmit....Send a correction file to BLS
4. Confirm.....Mark a correction as proper

Enter choice (#):

Each option is described in detail below.

List ... Create a Review Listing
AP option 1 submits a job that reads the State databases for the original input data
and any corrections (from the current annual processing period) on file. The data
are then displayed by year and month for each input variable. See the AP Output
section below for an example of the output.

Correct ... Modify a Month/Year Value for a Series
AP option 2 allows users to enter corrections/revisions to their State input data.
The input series for which modifications can be made will be displayed, as in the
example below:

LAUS Program Manual 12-21

STARS Annual Processing Module

Which series do you want to correct?
1. CES nonag w&s employment
2. CES adjustment for major strikes
3. UI continued claims w/o earnings
4. UCFE claims w/o earnings
5. Atypical UI claims adjustment
6. EXIT -- No more corrections

Enter choice (#):

For States with an approval to include extended unemployment
compensation (EUC) claims data, choice number 5 is Atypical UI claims
adjustment. This option will be displayed only for States with an
approved atypical.
Once a specific series is chosen, corrections/revisions for that series can
be made by entering data in the format shown. The entry must match the
format exactly, or it will not be accepted.
----- Corrections for CES nonag w&s employment -----

Enter data in this format: YY/MM/########

YY = 2-digit year, MM = 2-digit month, and ######## = data.
Press the Enter key at a new prompt when you are finished.

==>

92/12/423200

==>

92/13/425500
Invalid month.

***

==>

Entry not accepted

76/01/-4
Invalid year
Invalid data

12-22 LAUS Program Manual

***

STARS Annual Processing Module

***

Entry not accepted

***

==>

After corrections/revisions for one series are entered, another series (or the same
one) may be modified. Once all corrections/revisions are made, the user should
choose the “Exit -- No more corrections” option. This will submit a job to enter
the State input corrections and revisions into a database, where they will be stored
until the modifications are transmitted. (See the AP Transmit section below.) The
job also displays the data by year and month for each input variable. See the AP
Output section below for an example of the output.

AP Transmit... Create a Correction File and Send to BLS
After all input data corrections/revisions have been entered and checked carefully,
they must be officially transmitted to the national office. AP option 3 allows the
user to transmit the data; after transmittal, the data will be incorporated into the
official State databases. The transmit option appears on the annual processing
options screen, as shown in the example below. The user is prompted to make sure
that this is the desired option.
*

STARS Annual Processing Module

*

------------ Annual Processing options ----------1. List .......Create a review listing
2. Correct ....Modify a month/year value for a series
and create a correction file
3. Transmit ...Send a correction file to BLS
4. Confirm ....Mark a correction as proper
5. Exit
Enter choice (#):
***
***

3

Transmitting officially sends your
input changes to BLS files.

***
***

Are you sure you want to transmit?
(Y/n):

The transmit option should also be used after all confirmations (if any have been
requested by BLS) have been entered and checked carefully.

LAUS Program Manual 12-23

STARS Annual Processing Module

Confirm .... Mark a Correction as Proper
Sometimes, after input data corrections/revisions are transmitted, BLS
staff needs to confirm one or more of these corrections. AP option 4
allows users to enter confirmations of State input data corrections. The
input series for which confirmations can be made will be displayed, as in
the example below:

Which series do you want to confirm?
1.
2.
3.
4.
5.
6.

CES nonag w&s employment
CES adjustment for major strikes
UI continued claims w/o earnings
UCFE claims w/o earnings
Atypical UI claims adjustment
EXIT -- No more confirmations

Enter choice (#):

4

Once a specific series is chosen, confirmations for that series can be made
by entering data in the format shown. The entry must match the format
exactly, or it will not be accepted.
----

Confirmations for UCFE claims w/o earnings

Enter data in this format:

----

YY/MM/C

YY = 2-digit year, MM = 2-digit month (e.g., 93/01/C).
Press the Enter key at a new prompt when you are finished.
==>
==>

92/12/C
92/11/45
Invalid data.
***

Entry not accepted

***

==>

After entering confirmations for one series, data for another series (or the
same one) may be confirmed. Once all confirmations are made, the user
should choose the “Exit -- No more confirmations” option. This will
submit a job to enter the State input confirmations into a database, where
they will be stored until transmittal. Note that after all input data
confirmations have been entered and checked carefully, they must be
officially transmitted to BLS; see the AP Transmit section above. The job
also displays the data by year and month for each input variable. See the
AP Output section below for an example of the output.

12-24 LAUS Program Manual

STARS Annual Processing Module

AP Output
After the completion of an AP job, the output tables will be saved (on permanent
storage if a transmitted job, on temporary storage otherwise). These “table” files
should not be confused with the correction files which are created when using the
AP macro. A “correction” file is stored on permanent storage and modified each
time a correction is made. A separate file is created when corrections are
transmitted. As shown in the example below, the name of this file is displayed, and
the user can choose to list the output tables simply by pressing the Enter key.
The job is complete.
The output tables are in your active workspace and have
been saved as STARS.AP.D931215.T095534 on TMP.
Do you want to list the tables on your screen?
(Y/n):

A partial sample listing is shown below.
Annual Processing Review Listing
02/10/94
09:55:34
---------- Example ---------Data Series = CES nonag w&s employment

Year
92
92
92
92
92
92
92
92
92
92
92
92

Month

Correction
on File

1
2
3
4
5
6
7
8
9
10
11
12

.
.
.
365,700
.
.
.
.
.
.
417,500
423,200

Original
Data
381,200
382,700
385,400
395,200
402,600
417,400
420,600
426,200
416,100
415,600
412,000
411,500

Diff

% Diff

.
.
.
-29,500
.
.
.
.
.
.
5,500
11,700

.
.
.
-7.5
.
.
.
.
.
.
1.3
2.8

Confirm
.
.
.
Y
.
.
.
.
.
.
.
.

LAUS Program Manual 12-25

STARS Annual Processing Module

12-26 LAUS Program Manual

Glossary
"A" Factor The "A" factor imparts the influence of the experienced labor force

on that portion of the Handbook estimate covering total entrant unemployment.
Additional Claim An additional claim is 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 LMAs 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.
Annual Processing (AP) A series of activities conducted annually which

results 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 JTPA as

an area of at least 10,000 population with an average of 6.5 percent or more
unemployment in the previous 12 months. It is used for determining eligibility for
employment and training programs.
Autocorrelation This denotes that 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 p (rho) This coefficient is 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 strong positive autocorrelation.

LAUS Program Manual G-1

Glossary

"B" Factor The "B" factor imparts the influence of the experienced

unemployed on that portion of the Handbook estimate covering total
entrant unemployment.
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 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
amount.
Bias Bias is the difference between the expected value of the estimate

from a probability sample and the true value of the population parameter.
Bureau of Labor Statistics (BLS) Established in 1884 and now part

of the U.S. Department of Labor, this Federal agency functions as the
principal data-gathering agency of the Federal Government in the field of
labor economics. BLS collects, processes, analyzes, and disseminates
data relating to employment, unemployment, the labor force, productivity,
prices, family expenditures, wages, industrial relations, and occupational
safety and health.
Bureau of the Census (BOC) The BOC is a bureau of the U.S.

Department of Commerce. It conducts 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 Current Population Survey (CPS) for the
BLS.
Census A census is 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 This is a method used to disaggregate LMA employment

and unemployment estimates to subareas by assigning to the areas the
same proportion of the monthly, independent LMA estimate as was
evidenced in the most recent census.

G-2 LAUS Program Manual

Glossary

Census Tracts These Census-designated units are small parts of MSAs and

provide statistically 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.
Civilian 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 claim is 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 The claimant is 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 This is a method for

disaggregating LMA unemployment to subareas by using (1) claims by county of
residence to distribute Handbook experienced unemployment and (2) Census 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 These are 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
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) In statistics, this is the measure of relative

dispersion of data. The standard deviation divided by the arithmetic mean times
100 yields the coefficient of variation.

LAUS Program Manual G-3

Glossary

Commutation Regular travel of a person from the place of residence to

the job location or to the place of filing for UI benefits is referred to as
commutation.
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 Compositing is 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 This is a claim filed after the initial claim, by mail,

telephone, or in person, for waiting-period credit or payment for a
certified week of unemployment.
Correlation A statistical term which indicates a structural, functional, or

qualitative correspondence between comparable entities. Correlation can
be either positive (simultaneous increase or decrease in both variables) or
negative (increase in the value of one and decrease in the value of the
other variable).
Correlation Coefficient This coefficient is 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 strong positive
relationship. A value of "-1" indicates strong negative relationship,
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 selfemployed workers.
Current Employment Statistics (CES) program The CES is a

monthly sample survey of about 400,000 employers which yields
estimates of nonagricultural wage and salary employment, hours, and
earnings by industry. These statistics are prepared monthly by the BLS

G-4 LAUS Program Manual

Glossary

for the nation as a whole, and by cooperating State agencies for each of the 50
States, the District of Columbia, and most MSAs. The BLS published CES data
in "Employment and Earnings."
Current Population Survey (CPS) The CPS is a monthly survey conducted

by the Bureau of the Census of approximately 55,000 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 A denial is 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 dependent variable is 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.
Disaggregation Disaggregation divides a statistic into its component parts. For

example, the LMA unemployment is divided into each component county or city.
Discouraged Workers Persons not in the labor force who want and are

available for a job and who have looked for work some time 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.
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 are earnings disregarded. This
is also referred to as the forgiveness level for earnings. The amounts vary by
State.
Earnings Due to Employment These are any 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.

LAUS Program Manual G-5

Glossary

Employment/Population Ratio The civilian employment/population

ratio is the proportion of the civilian noninstitutional population who are
classified as employed. The State E/P ratio is the dependent variable in
the equation which yields monthly State employment.
Employment and Training Administration (ETA) An agency within

the Department of Labor which includes the Office of Job Training, the
U.S. Employment Service, and the Unemployment Insurance Service.
Enumeration Districts (EDs) Administrative units used in the Census

are referred to as enumeration districts. 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 establishment is defined as 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 Persons 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) This is 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.
Extrapolate 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 are issued by the national Bureau of Standards in
the U.S. Department of Commerce. They include a geographically
exhaustive five digit code system wherein areas such as State, counties,
territories, and MSAs are uniquely identified.
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.

G-6 LAUS Program Manual

Glossary

Gain The Gain is 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.
Handbook Method This is a technique used to provide the basis for the monthly

estimate of total unemployment and total employment at the labor market area
level. This method, referred to as the Handbook Method because of its inclusion
in the earlier "Handbook on Estimating Unemployment", is a building block
approach using certain key statistics on the insured unemployment and on
nonfarm payroll employment which are available on a monthly basis and for all
areas in the State.
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 ICON (Interstate Connection) is a centralized computerized system of

reporting and exchanging unemployment insurance claims information between
States.
Independent Variables These are variables used in the regression equation to

predict the dependent variable, "Y". The independent variables are usually termed
the "X" variables.
Initial Claim An initial claim is any notice of unemployment filed by an

individual to initiate (1) a determination of entitlement to an d 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 The institutional population is comprised of persons

residing in the following types of institutions: penal institutions, mental
institutions, sanitaria, homes for the aged or inform, and homes for the needy.
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.

LAUS Program Manual G-7

Glossary

Intercept The value of "Y" (dependent variable) where the regression

line crosses the "Y" axis is the intercept. The intercept is usually denoted
by βo.
Interpolate To estimate values of a function between two known values.
Interstate claim An interstate claim is 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 An intrastate claim is 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 Job leavers are persons who quit or otherwise terminate

their employment voluntarily and immediately begin looking for work.
Job Losers Job losers are persons on layoff and those whose

employment ended involuntarily and who immediately begin looking for
work.
Kalman Filter The Kalman Filter is 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 civilian labor force comprises 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) A LMA is an economically integrated

geographical unit within which workers may readily change jobs without
changing their place of residence. All States are divided into exhaustive
LMAs, which include a county or a group of contiguous counties, except
in New England where cities and towns are used. Independent Handbook
estimates of employment and unemployment are made monthly for each
LMA and form the basis for the LAUS estimates.
Labor Surplus Area This is 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%.)

G-8 LAUS Program Manual

Glossary

LAUS Estimate These are 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.
Least Squares Least Squares is a basic regression 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 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 The link relative technique for employment

estimation 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 JTPA,
FEMA, etc.
Mass Layoff Event This is a layoff in which 50 initial claims or more have been

filed against an establishment during a five-week period, with the separations
expected to last longer than 30 days.
Mean Square Error (MSE) MSE is 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 is variance and is an
important statistical and analytical tool.
Metropolitan Statistical Area (MSA) An MSA is a geographic area

comprised of a county generally containing a central city (or twin cities) of 50,000
inhabitants or more, plus contiguous counties that are socially and economically
integrated with the central city. (New England MSAs are comprised of towns and
cities, rather than counties.)
Migration Migration is the permanent movement of an individual's residence

from one location to another.

LAUS Program Manual G-9

Glossary

Model A model is a mathematical equation which relates different

variables and data. In time series, this relationship is computed over time.
In the LAUS signal-plus-noise models, the monthly State CPS labor force
estimates are related to different independent variables and data that show
strong correlations to the monthly estimates.
Monetary Determination A written notice is 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.
Months for Cyclical Dominance (MCD) The MCD is an estimate of

the time span required to identify significant cyclical movements in a
monthly economic time series. The MCD indicates the shortest span of
months over which changes in the series are dominated by cyclical rather
than irregular or erratic movements.
Moving Average This is 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 A new claim is the first initial claim filed in person, by mail,

or telephone to request a determination of entitlement to and eligibility for
compensation. This is one of three types of initial claims.
New Entrants In the CPS, new entrants are new workers looking for a

job. 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.

G-10 LAUS Program Manual

Glossary

Nonmonetary Determination This process 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 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 Residency Adjustment Factor.
Population-Based Indexed Share Employment Disaggregation This

method 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 is used only in conjunction with the claims-based unemployment
disaggregation.
Population Estimates The Bureau of the Census annually prepares total

population estimates for States and selected substate areas.
Population Share The population share method is used for disaggregating an

area's employment and unemployment estimates to places by assigning to the
place the same proportion of employment and unemployment as its proportion of
the larger area's census population. It is used in lieu of census share where census
data on employment and unemployment do not exist for an area.
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 The prediction period is 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 Metropolitan Statistical Area (PMSA) A PMSA is a subarea

defined within an area that meets the requirements to qualify as an MSA and also
has a population of one million or more. Upon the designation of PMSAs, the
entire area of which they are parts becomes a Consolitated Metropolitan Statistical
Area.

LAUS Program Manual G-11

Glossary

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.
Quarterly Report of Employment, Wages, and Contributions
(ES-202 Report) Employment and wage data for workers covered by

State unemployment insurance laws and civilian workers covered by
UCFE comprise the ES-202. It is compiled from quarterly tax reports
submitted to State Employment Security Agencies by employers. The
Quarterly Report is referred to as the Es-202 because this is the
identifying number of the Federal report form which summarizes the data
from quarterly tax reports.
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 interative basis in the CPS.
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 the calendar week including the 12th of the month.
For UI data, it is the certification period. In most States, the reference
week for UI certifications is the calendar week including the 12th.
Exceptions are States with flexible benefit weeks and New York, whose
week is a Monday-through-Sunday week.
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 Equations The basic formula for a regression
equation is shown below. In this example, the equation has an
intercept, Bo, two independent variables, X1 and X2, with
coefficients B1 and B2, respectively. The equation's error term is E.

Y = Bo +B1 * X1 + B2 * X2 + E

G-12 LAUS Program Manual

Glossary

Regression Line A regression line is 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.
Residency Adjustment Factor In order to convert the CES nonfarm job count

to a person count by place of residence, the census estimate for an area is used as a
base which is then extrapolated by the percent change in the establishment-based
employment series. This is mathematically equivalent to multiplying the current
establishment-based employment estimate by the ratio of the census to the
establishment-based employment estimate for the census reference period. It is
derived and used in the preparation of the current Handbook estimates.
Rotation Group A rotation group is one of eight systematic subsamples which

comprise the total CPS sample. A rotation group is in the sample for four
consecutive months 1 year, leaves the sample during the following eight months,
and then returns for the same four calendar months of the next year.
Sample A sample is a subset of a statistical population selected for the purpose

of making generalized statements about the whole.
Sample Period The sample period is a period of time which is used to derive

coefficients for the regression line. It is also called the "inside sample" period.
Sampling Error Sampling error is 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 This is 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 is the sampling ratio. Sampling ratios
vary by cell, depending on the degree of variability of the measured item.
Scatter Plot This is 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.
Seasonal Adjustment Seasonal adjustment of time-series data is done to

eliminate the effect of intra-year variations which tend to occur each year in
approximately the same manner. Examples of such variations include school
terms, holidays, weather patterns, etc.

LAUS Program Manual G-13

Glossary

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.
Signal-Plus-Noise Models These are 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 The slope is 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".
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-back-forward
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.

Smoothing

Standard Deviation Standard deviation is a measure of dispersion

around the mean value of a population frequently denoted by sigma (s). 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.
Standard Industrial Classification (SIC) This is the statistical

classification standard underlying all establishment-based Federal
economic statistics identified by industry. The SIC is used to promote the
comparability of establishment data describing various facets of the
United States economy.
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

G-14 LAUS Program Manual

Glossary

which is an exchange for workers seeking work and employers seeking workers.
(3) Research and Analysis which includes collection, analysis, and publication of
labor market information.
Statistical Population This is 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 Stochastic is a term often 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 time series is 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.
Transitional Claim A transitional claim is 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.
Unemployment Compensation for (E)X-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 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 Unemployed disqualified refers to 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.

LAUS Program Manual G-15

Glossary

Unemployment Rate The number of persons unemployed, expressed

as a percentage of the civilian labor force.
Variable A variable is 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 rate and CES employment.
Variable Coefficient Model (VCM) The VCM is a type of sample

regression model in which the model's coefficients are allowed to change
over time.

Variance is 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 (s2).
Variance

G-16 LAUS Program Manual

Index
A
A and B factors 7-20
regression formulas 7-20
additivity 8-1–8-5
handbook share method 8-3–8-4
handbook share method example 8-4
interstate areas 8-5
Agriculture, Department of, 1-22
agricultural unemployment 7-19
annual model review 10-1
annual processing 10-1
smoothing 10-7
annual re-estimation 10-7
Kalman Filter 10-7
Appalachian Regional Commission 1-25
autocorrelation 6-6, 6-7

B
benchmarking 10-7
changes in geographical areas 10-14
incorporation of revised sub-State
data updates 10-13
State estimates 10-9
sub-State estimates 10-12
sub-State processing 10-15
The Denton Method 10-9-10-11
Bureau of the Census 2-1

C
Census 1-10, 5-1
census versus CPS/LAUS 5-4
estimation 5-2
nonsampling errors 5-3
questionnaire 5-1
statewide estimation 5-9
substate estimation 5-10
administrative records method 5-10
component method 5-10

regression method 5-10
use of census data in LAUS 5-6
uses of labor force estimates 5-6
uses of population data 5-6, 5-7
census-share method 9-1
CES employment 1-10
civilian noninstitutional population
(CNP) 2-4

population projections 2-26
claims taking 3-3
claims-based unemployment 9-7
Commerce, Department of, 1-18
computer assisted personal interviewing
(CAPI), 2-2
computer assisted telephone interviewing
(CATI) 2-2
consolidated metropolitan statistical areas
(CMSA) 7-3
continued claimants 3-9
continued claims without earnings 3-10
Covered Employment and Wages (ES202) 4-5
agents on commission 4-8
agriculture 4-7
data compilation 4-6
data uses 4-9
domestic service 4-7
earnings data 4-8
military personnel 4-8
nonprofit organization 4-7
railroads 4-7
self-employed and unpaid family 4-7
sources of data 4-5
state and local government 4-8
student workers 4-8
unpaid absences 4-6
Current Employment Statistics (CES) 4-1
benchmarks 4-3
employment defined 4-2
establishment defined 4-1
estimating current LMA employment 4-12

LAUS Program Manual i-1

Index

estimation process 4-2
reliability 4-4
residency-adjustment 4-13
state employment models 4-12
Current Population Survey (CPS) 2-1
assigned households. 2-15-2-16, 2-20
basic weighting 2-23
coefficient of variation (CV) 2-10
composite estimation 2-26
computer assisted personal interview (CAPI) 2-21
computer assisted telephone interview (CATI) 2-20
concepts and definitions 2-4–2-7
designated households 2-20
eligible households 2-20
estimation 2-22–2-26
first-stage ratio 2-23
noninterview 2-24
nonresponse 2-9
nonsampling error 2-8
personal visits 2-20
raking process - see second stage ratio 2-24
reference week 2-3, 2-20
rotation groups 2-17
sample bias 2-18
sample design 2-12–2-18
sample overlap 2-17–2-18
sample rotation design 2-17
sampling error 2-10
seasonal adjustment 2-26
second-stage ratio 2-24
standard error 2-8
state sample sizes 2-15

D
data inputs to LAUS estimates 1-9
Defense, Department of, 1-23
disaggregation 9-1-9-20
census-share method 9-1, 9-19
claims data 9-9
claims-based unemployment 9-9, 9-16
employment disaggregation procedure 9-4
experienced unemployed (geographic basis
of claims data) 9-16

LAUS Progam Manual i-2

overview 9-1
population-based employment 9-15
population-claims method 9-3
population-claims method for counties 9-3
population-claims method for

incorporated places 9-14
population-claims of interstate areas 9-13
specification of population size 9-14
unemployment disaggregation procedure 9-10
unemployment procedure for cities or towns 9-17
discouraged workers 2-4
duration of unemployment 2-7

E
employed persons 2-5
employment model 6-10–6-11
CES variable 6-10
time series components 6-11
enumerative check census 2-1
ES-202 1-10
establishment data sources versus the CPS 4-10
age of workers 4-10
employment coverage 4-10
jobs versus employed people 4-10
place of work vs. residence. 4-10
reference periods 4-10
unpaid absence 4-10
estimation methods 1-11
exhaustees 7-15-7-17
Extract Macro 11-1

F
Federal Unemployment Tax Act (FUTA) 4-7
Federal Emergency Management Agency 1-21
final payment recipient 3-9

H
handbook estimation 7-4
agricultural employment 7-10
agricultural employment benchmark 7-11
agricultural estimating procedure 7-13

Index

agricultural estimating regions 7-11
Agricultural Labor Survey (ALS) 7-10
agricultural unemployment 7-17
all-other employment 7-7
area strata development 7-8
covered unemployment 7-14
k value 7-8
labor market area employment 7-6
labor market area unemployment 7-14
new and reentrant unemployed 7-18
nonagricultural wage and salary (NAWS)
employment 7-6
noncovered unemployment 7-18
step 3 ratio 7-9
unemployed exhaustees 7-15
Health and Human Services, Department of 1-24
history 1-3–1-8

I
initial claim 3-10
exclusions 3-11

Levitan Commission 6-1

M
metropolitan statistical areas (MSA) 4-1, 7-2
monetary determination 3-3
multiple jobholders 2-5

N
National Center for Health Statistics (NCHS) 2-25
new and reentrant unemployment 7-19
New England county metropolitan areas (NECMA) 7-4
new entrant unemployed 7-19
new entrants 2-6
noncovered unemployment 7-18
nonmonetary determination 3-3
disqualification 3-11
penalties 3-13
unemployed disqualified 3-11
Not in the Labor Force 2-4

O
J
Job leavers 2-6
job losers 2-6
Justice, Department of, 1-23

K
Kalman Filter 6-2, 6-16–6-17

L
Labor, Department of, 1-19
labor force 2-4
labor market area (LMA) estimation 7-1-7-40
defined 7-1
defining small labor maket areas 7-3-7-4
overview 7-1
labor market area unemployment 7-14
labor market level estimation 1-11
LAUS publications 1-24

Office of Personnel Management 1-27
outliers 6-8–6-9

P
population controls 10-3
independent population controls 2-9
State annual controls 10-5
sub-State annual controls 10-6
population estimates 10-3
administrative records 10-4
components-of-change 10-4
population-claims method 9-1
Post Enumeration Survey 2-25
primary metropolitan statistical areas (PMSA) 7-3
primary sampling units (PSU) 2-2
defined 2-12
non-self-representing 2-13
probability of selection 2-13-2-14, 2-17
self-representing 2-13

LAUS Program Manual I-3

Index

stratification 2-13

R
reentrant unemployed 7-19
re-entrants 2-6
reference period 3-9
residency coding 3-9

S
seasonal monthly A and B factors 7-21-7-22
self-employed 2-5
signal-plus-noise models 6-3–6-9
correlated sampling error 6-5
CPS sample overlap 6-5–6-8
CPS sample rotation 6-5–6-8
CPS sampling error 6-4
data revision 6-17–6-18
error term 6-15
estimation process 6-15–6-16
formulas 6-4
seasonal adjustment 6-19–6-21
signal components 6-14–6-15
time series parameter estimation 6-14
Social Security Administration 2-26
standard error 2-10
Standard Industrial Classification (SIC) 4-6
STARS 11-1–11-16
annual processing 12-21
BLS notices 11-2
CES employment 11-1, 12-4, 12-5
changes from prior month/year 12-8
coefficients 11-12
components of change 11-10, 12-8
components of signal 11-12, 12-8
continued claims without earnings 12-4, 12-5
CPS 11-1
create new estimates 12-3
data extract output 12-12
entering data 12-5
estimation 12-7
execution 12-1

Extract Macro 11-1
input data 11-13, 12-5, 12-8
main options 12-3
major strikes 12-4, 12-5
monthly tables 12-8
official transmission 12-19
over-the-month and over-the-year changes 11-4, 11-5
plots 11-14, 12-8
primary functions 11-1
regression explanatory variable 11-12
residual seasonal 11-12
residual trend 11-12
review module 12-10–12-18
state comments 11-2
strikers 11-2
SunGard 12-1
SunGard access 12-1
text file 12-14
transmit current estimates 12-19
UCFE claims 12-5
unemployment claims 11-1
user’s guide 12-1
variable descriptions 12-16
variables 12-11
year-to-date model estimates 11-4, 12-8
STARS user’s guide 12-1–12-26
state level estimation 1-11
State unemployment insurance (UI) 3-1
strikers 11-2
SunGard 12-1
communication settings 12-1
logon steps 12-1

T
temporary jobs 2-6
Treasury, Department of, 1-25

U
UCFE claims 12-4

LAUS Progam Manual i-4

Index

UI 1-9
UI database survey 3-6
ultimate sampling units (USU) 2-12
defined 2-14
unemployed entrant estimation 7-19
unemployed exhaustee worksheet 7-17
unemployed persons 2-6
unemployed reentrant estimation 7-18
unemployment claims 11-2
Unemployment Compensation for Federal Employees
(UCFE) 3-1
unemployment insurance (UI) system 3-1-3-14
exclusions 3-11
overview 3-1
state responsibility 3-3
unemployment rate 2-7
unemployment rate model 6-12–6-13
components 6-13
UI claims variable 6-12
unpaid family worker 2-5
uses/users of LAUS data 1-13

V
variable coefficient model (VCM) 6-2

W
wage and salary workers 2-5
Work Projects Administration (WPA) 2-1

X
X-11 ARIMA 2-2, 2-27, 6-19, 6-20
model 2-2

Y
youth population ratio (YPR) 7-20

LAUS Program Manual I-5