Technical Note
This news release presents civilian labor force and unemployment data for states
and selected substate areas from the Local Area Unemployment Statistics (LAUS)
program (tables 1 and 2). Also presented are nonfarm payroll employment estimates
by state and industry supersector from the Current Employment Statistics (CES)
program (tables 3 and 4). The LAUS and CES programs are both federal-state cooperative
endeavors.
Civilian labor force and unemployment--from the LAUS program
Definitions. The civilian labor force and unemployment data are based on the same
concepts and definitions as those used for the official national estimates obtained
from the Current Population Survey (CPS), a sample survey of households that is
conducted for the Bureau of Labor Statistics (BLS) by the U.S. Census Bureau.
The LAUS program measures employed people and unemployed people on a place-of-
residence basis. The universe for each is the civilian noninstitutional population
16 years of age and older. Employed people are those who did any work at all for
pay or profit in the reference week (typically the week including the 12th of the
month) or worked 15 hours or more without pay in a family business or farm, plus
those not working who had a job from which they were temporarily absent, whether
or not paid, for such reasons as bad weather, labor-management dispute, illness,
or vacation.
Unemployed people are those who were not employed during the reference week
(based on the definition above), had actively looked for a job sometime in the
4-week period ending with the reference week, and were currently available for
work; people on layoff expecting recall need not be looking for work to be counted
as unemployed. The civilian labor force is the sum of employed and unemployed people.
The unemployment rate is the number of unemployed as a percent of the civilian labor
force.
Method of estimation. Estimates for 48 states, the District of Columbia, the Los
Angeles-Long Beach-Glendale metropolitan division, New York City, and the balances
of California and New York State are produced using time-series models. This method,
which underwent substantial enhancement at the beginning of 2015, utilizes data from
several sources, including the CPS, the CES, and state unemployment insurance (UI)
programs. Estimates for the state of California are derived by summing the estimates
for the Los Angeles-Long Beach-Glendale metropolitan division and the balance of
California. Similarly, estimates for New York State are derived by summing the estimates
for New York City and the balance of New York State. Estimates for the five additional
substate areas contained in this release (the Cleveland-Elyria and Detroit-Warren-
Dearborn metropolitan areas and the Chicago-Naperville-Arlington Heights, Miami-
Miami Beach-Kendall, and Seattle-Bellevue-Everett metropolitan divisions) and their
respective balances of state are produced using a similar model-based approach.
Each month, estimates for the nine census divisions first are modeled using inputs
from the CPS only and controlled to the national totals. State estimates then are
controlled to their respective census division totals. Substate and balance-of-state
estimates for the five areas noted above also are controlled to their respective state
totals. This tiered process of controlling model-based estimates to the U.S. totals is
called real-time benchmarking. Estimates for Puerto Rico are derived from a monthly
household survey similar to the CPS. A more detailed description of the estimation
procedures is available from BLS upon request.
Annual revisions. Civilian labor force and unemployment data for prior years reflect
adjustments made after the end of each year. The adjusted estimates reflect updated
population data from the U.S. Census Bureau, any revisions in the other data sources,
and model re-estimation. In most years, historical data for the most recent five years
are revised near the beginning of each calendar year, prior to the release of January
estimates. With the introduction of a new generation of times-series models in early
2015, historical data were re-estimated back to the series beginnings in 1976, 1990,
or 1994.
Seasonal adjustment. The LAUS models decompose the estimates of employed and unemployed
people into trend, seasonal, and irregular components. Prior to 2018, the benchmarked
trend component of each measure had been smoothed using a Trend-Cycle Cascade Filter.
With changes implemented in early 2018, the benchmarked signals of employed and
unemployed people first are adjusted using an X-11 type of seasonal adjustment filter.
The adjusted data then are smoothed using a Reproducing Kernel Hilbert Space (RKHS)
filter. The smoothed-seasonally adjusted estimates of employed and unemployed people
are summed to derive the civilian labor force, and the unemployment rate then is
calculated as the unemployed percent of the civilian labor force. The resulting
smoothed-seasonally adjusted unemployment rate estimates are analyzed in this news
release and published on the BLS website.
During estimation for the current year, the smoothed-seasonally adjusted estimates
for a given month are created using an asymmetric filter that incorporates information
from previous observations only. For annual revisions, historical data are smoothed
using a two-sided filter.
In early 2018, historical data were re-estimated back to the series beginnings in
1976, 1990, or 1994 to incorporate the changes to the seasonal adjustment and smoothing
procedures described above.
Area definitions. The substate area data published in this release reflect the
delineations that were issued by the U.S. Office of Management and Budget on July 15,
2015. A detailed list of the geographic definitions is available online at
www.bls.gov/lau/lausmsa.htm.
Employment--from the CES program
Definitions. Employment data refer to persons on establishment payrolls who receive
pay for any part of the pay period that includes the 12th of the month. Persons are
counted at their place of work rather than at their place of residence; those appearing
on more than one payroll are counted on each payroll. Industries are classified on
the basis of their principal activity in accordance with the 2017 version of the North
American Industry Classification System.
Method of estimation. CES State and Area employment data are produced using several
estimation procedures. Where possible these data are produced using a "weighted link
relative" estimation technique in which a ratio of current month weighted employment
to that of the previous-month weighted employment is computed from a sample of
establishments reporting for both months. The estimates of employment for the current
month are then obtained by multiplying these ratios by the previous month's employment
estimates. The weighted link relative technique is utilized for data series where
the sample size meets certain statistical criteria.
For some employment series, the sample of establishments is very small or highly
variable. In these cases, a model-based approach is used in estimation. These models
use the direct sample estimates (described above), combined with forecasts of
historical (benchmarked) data to decrease volatility in estimation. Two different
models (Fay-Herriot Model and Small Domain Model) are used depending on the industry
level being estimated.For more detailed information about each model, refer to
the BLS Handbook of Methods.
Annual revisions. Employment estimates are adjusted annually to a complete count
of jobs, called benchmarks, derived principally from tax reports that are submitted
by employers who are covered under state unemployment insurance (UI) laws. The
benchmark information is used to adjust the monthly estimates between the new
benchmark and the preceding one and also to establish the level of employment
for the new benchmark month. Thus, the benchmarking process establishes the
level of employment, and the sample is used to measure the month-to-month
changes in the level for the subsequent months. Information on recent benchmark
revisions is available online at www.bls.gov/web/laus/benchmark.pdf.
Seasonal adjustment. Payroll employment data are seasonally adjusted at the
statewide expanded supersector level. In some cases, the seasonally adjusted
payroll employment total is computed by aggregating the independently adjusted
supersector series. In other cases, the seasonally adjusted payroll employment
total is independently adjusted. Revisions to historical data for the most recent
five years are made once a year, coincident with annual benchmark adjustments.
Beginning with the release of January 2018 preliminary estimates, payroll employment
data are seasonally adjusted concurrently, using all available estimates including
those for the current month, to develop sample-based seasonal factors. Concurrent
sample-based factors are created every month for the current month's preliminary
estimate as well as the previous month's final estimate in order to incorporate
real-time estimates. Previously, the sample-based seasonal factors were forecasted
once annually at the beginning of the year and applied to the sample-based estimates
for the 12 months of the year.
Caution on aggregating state data. State estimation procedures are designed to
produce accurate data for each individual state. BLS independently develops a
national employment series; state estimates are not forced to sum to national
totals. Because each state series is subject to larger sampling and nonsampling
errors than the national series, summing them cumulates individual state-level
errors and can cause significant distortions at an aggregate level. Due to
these statistical limitations, BLS does not compile a "sum-of-states" employment
series, and cautions users that such a series is subject to a relatively large
and volatile error structure.
Reliability of the estimates
The estimates presented in this release are based on sample surveys, administrative
data, and modeling and, thus, are subject to sampling and other types of errors.
Sampling error is a measure of sampling variability--that is, variation that occurs
by chance because a sample rather than the entire population is surveyed. Survey
data also are subject to nonsampling errors, such as those which can be introduced
into the data collection and processing operations. Estimates not directly derived
from sample surveys are subject to additional errors resulting from the specific
estimation processes used.
Use of error measures. Changes in state unemployment rates and state nonfarm payroll
employment are cited in the analysis of this release only if they have been determined
to be statistically significant at the 90-percent confidence level. Furthermore, state
unemployment rates for the current month generally are cited only if they have been
determined to be significantly different from the U.S. rate at the 90-percent confidence
level. The underlying model-based standard error measures for unemployment rates
and over-the-month and over-the-year changes in rates are available at
www.bls.gov/lau/lastderr.htm. The underlying standard error measures for over-the-month
and over-the-year changes in state payroll employment data at the total nonfarm and
supersector levels are available at www.bls.gov/web/laus/790stderr.htm. Measures of
nonsampling error are not available.
Additional information
Estimates of civilian labor force and unemployment from the LAUS program, as well
as nonfarm payroll employment from the CES program, for metropolitan areas and
metropolitan divisions are available in the news release Metropolitan Area Employment
and Unemployment. Estimates of civilian labor force, employed people, unemployed
people, and unemployment rates for approximately 7,000 subnational areas are
available online at www.bls.gov/lau/. Employment data from the CES program for
states and metropolitan areas are available online at www.bls.gov/sae/. Information
in this release will be made available to sensory impaired individuals upon request.
Voice phone: (202) 691-5200; Federal Relay Service: (800) 877-8339.
United States Department of Labor
