# Chapter 4

Measurement of Unemployment in States and Local Areas (PDF)

Estimates of unemployment in states and local areas are
key indicators of local economic conditions. These
estimates are produced by the Local Area Unemployment Statistics (LAUS) program and 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. They also are used by economic
forecasters, researchers, and bond and mortgage underwriters. In addition, local area unemployment estimates are used
to determine the eligibility of an area for benefits in various
federal assistance programs.

The LAUS program is a Federal-State cooperative program. The Bureau of Labor Statistics (BLS) develops concepts, definitions, and technical procedures and then works
with state workforce agencies, who prepare labor force and
unemployment estimates. Monthly and annual average estimates of employment and unemployment are prepared in
state agencies for more than 7,300 unique geographic areas:
states, the District of Columbia, and Puerto Rico; labor market areas (LMAs), such as metropolitan and micropolitan
areas; counties and equivalents; cities with a population of
25,000 or more; and all cities and towns in New England,
regardless of population.

### Background and History

Unemployment estimates have been developed for subnational areas for nearly 70 years. The program began during
World War II under the War Manpower Commission, with
the aim of identifying areas where a labor market imbalance
had been created due to inadequate labor supply, material
shortages, or transportation difficulties. After the war, emphasis was placed on identifying areas of labor surplus, and
a program of classifying areas in accordance with severity of
unemployment was established.

In 1950, the Department of Labor's Bureau of Employment Security (now Employment and Training Administration) published a handbook, Techniques for Estimating
Unemployment, so that comparable estimates of the unemployment rate could be produced for all states. This led to the
formulation of the "handbook method" in the late 1950s. The
handbook method is a series of computational steps designed
to produce local employment and unemployment estimates,
using available data at a much lower cost than a direct survey.
The handbook method relies heavily on data derived from
the unemployment insurance (UI) system. (See section "Estimates for Substate Areas—The handbook method.")

In 1972, BLS assumed technical responsibility for the
program and began to refine the concepts and methods used
to estimate the labor force, employment, and unemployment
at the subnational level. In 1973, a new system for developing labor force estimates was introduced, combining the
handbook method with the concepts, definitions, and estimation controls from the Current Population Survey (CPS).
The CPS, a monthly BLS survey conducted by the Census
Bureau, is used to measure the labor force status of individuals for the nation as a whole. (See chapter 1.) CPS estimates are based on data from a sample of households that is
designed to provide reliable monthly unemployment estimates for the nation and reliable annual average estimates for
the 50 states and the District of Columbia. A major advantage
of the CPS is that it applies uniformly across states, whereas
UI data are affected by individual states' UI laws.

Since 1973, the CPS sample size has been increased (and
reduced to a lesser degree) several times in selected states, and
the design has been modified to improve the quality of state
labor force estimates. As a criterion for using the monthly
CPS data directly for official publication of labor force estimates, a maximum expected coefficient of variation (CV) of 10 percent for unemployment—assuming an unemployment
rate of 6 percentâ€”was established by BLS. (The coefficient
of variation of an estimate is defined as the standard error of
the estimate divided by the estimate itself.) On the basis of
this criterion, beginning in 1978 the monthly CPS data were
used for official statewide labor force estimates for 10 large
states—California, Florida, Illinois, Massachusetts, Michigan,
New Jersey, New York, Ohio, Pennsylvania, and Texas—and
for 2 substate areas—the Los Angeles-Long Beach metropolitan area and New York City. These states and areas were referred to as "direct-use" areas, because they used the CPS data
without any mathematical or statistical adjustments. Official
monthly estimates for the remaining "non-direct-use" states
were based on the handbook method adjusted to CPS controls,
using a historical 6-month moving-average ratio adjustment

In 1985, a state-based design for the CPS was fully implemented for the first time, to incorporate 1980 Census information and to improve reliability for each of the 50 states
and the District of Columbia. North Carolina was added as
another direct-use state, and the CV requirement for monthly
unemployment was reduced to 8 percent for these 11 large
states. For each of the other 39 (non-direct-use) states and the
District of Columbia, the reliability requirement was established at an 8-percent CV for annual average unemployment,
assuming a 6-percent unemployment rate.

Beginning in 1989, handbook estimation for the 39 nondirect-use states and the District of Columbia was discontinued in favor of time-series statistical modeling. The models
were developed by BLS and tested by state workforce agencies. (Estimates for most substate areas continue to be based
on the handbook method.)

Estimates for the large, direct-use states had been seasonally adjusted since 1978. In 1992, seasonal adjustment was
extended to the model-based estimates for the non-direct-use
states. In 1994, in conjunction with a major redesign of the
CPS, a second generation of time-series models was introduced, based on a "signal-plus-noise" approach.

In 1996, in response to budget reductions, the number of
households in the CPS sample was temporarily reduced, resulting in the elimination of direct use of the CPS for monthly
estimation in the 11 large states, the Los Angeles-Long Beach
metropolitan area, and New York City. Beginning with January 1996, labor force estimates for these subnational areas
have been based on the time-series modeling approach used
in the other 39 states and the District of Columbia.

In 2005, improved third-generation time-series estimates
for modeled areas were introduced, along with real-time
benchmarking of state estimates to the national CPS estimates. (See the section "Estimates for States—Real-time Benchmarking.") Also introduced were new time-series
models for five metropolitan areas and the respective balances of their states, as well as improved substate estimation
in the handbook method.

In 2010, a process for smoothing the seasonally adjusted
estimates was introduced for all model-based areas. In 2011,
seasonal adjustment, including the smoothing process, was
introduced for metropolitan areas and metropolitan divisions.

### Estimation Methodology

#### Estimates for states

Monthly labor force data for all states and the District of
Columbia are based on the time-series approach to sample
survey data (Scott, Smith, and Jones, 1974; Bell and Hillmer, 1990). This approach reduces the high variability in
monthly CPS estimates that results from these areas' small
CPS sample sizes. Actual monthly CPS sample estimates
are represented in signal-plus-noise form as the sum of a
stochastic true labor force series (signal) and error (noise)
generated by sampling only a portion of the total population, where:

*y*_{t} = *Y*_{t} + *e*_{t}

*y*_{t} = CPS estimate,

*Y*_{t} = true labor force value (signal), and

*e*_{t} = sampling error (noise).

The signal is represented by a time-series model that incorporates historical relationships in the monthly CPS estimates, along with auxiliary data from the UI and Current
Employment Statistics (CES) programs. (See chapter 2.) This
time-series model is combined with a noise model that reflects key characteristics of the sampling error (SE), to produce estimates of the true labor force values (Tiller, 2006).
This estimator is optimal under the model assumptions and
has been shown by Bell and Hillmer (1990) to be design consistent under general conditions.

Two models—one for the unemployment level and a second for the employment level—were developed for each
state, on the basis of data from 1976 forward. The labor force
level and unemployment rate are derived from the employment and unemployment measures produced by the two respective models. The signals for both models are based on a
basic structural model that decomposes a series into stochastic trend-cycle (*T*_{t}), seasonal (*S*_{t}), and irregular (*I*_{t}) components (Harvey, 1989). The model is of the form:

*Y*_{t} = *T*_{t} + *S*_{t} + *I*_{t}

The trend-cycle and seasonal components have mutually
independent normal disturbance terms that cause them to drift
slowly over time. The variances of these disturbances constitute the hyperparameters of the signal and determine the
properties of the individual components. A positive variance
for a component implies that it is stochastic (not perfectly
predictable from past history), while a zero variance implies
deterministic behavior (a fixed pattern over time). The irregular is treated as an uncorrelated zero-mean disturbance with
fixed variance.

The model uses information from state CPS time series.
A natural extension of the structural model is to allow one or more of the unobserved components of the signal to be
related to corresponding components in another series. A
common core of state-specific monthly covariates have been
developed from auxiliary data sources: UI claims from the
Federal-State UI system are used for the unemployment
model, and nonagricultural payroll employment estimates
from the CES program are used for the employment model.
The model for the covariate, *X*_{t}, follows the same basic structural form as for *Y*_{t}, with stochastic trend, seasonal, and irregular components. The two series, *Y*_{t} and *X*_{t}, are treated as
related in a bivariate time-series model with contemporaneous correlations between their respective trend disturbances
(a special case of the seemingly unrelated time-series equations model; see Harvey, 1989). Correlations between irregular and seasonal components could be allowed, but because
they are very weak, adding further complexity would result
in little gain.

The bivariate model just described can accommodate a
wide variety of evolving CPS time-series patterns and covariate relationships. The degree to which the time-series components vary over time is determined empirically for each
state. In some cases, the seasonal component is estimated to
have a fixed pattern from year to year. For most models, the
irregular component is zero. Also, the degree of correlation
between the trends in the CPS and the covariates is determined empirically. The strongest relationship occurs when
there is a linear relationship between the trend levels and/or
growth rates (cointegration). For most of the CPS series, empirical correlations are not strong enough to imply the presence of cointegration with the covariates. For some series, the
trend correlations are effectively zero.

Occasionally, there are sudden changes, either temporary
or permanent, in the CPS that are not predictable from past
history. These aberrant observations, or outliers, are modeled
with exogenous regression variables that introduce independent outlier components into the model's components. The
most common types of outliers are permanent shifts and transitory changes in the level of the series. For example, a level
shift was introduced into the Louisiana models to account for
Hurricane Katrina.

The second major component of the signal-plus-noise
model deals with CPS sampling errors (Tiller, 1992). Because of this survey's complex design, the behavior of the
observed sample estimates differs in important ways from
that of the true values. Sampled households are rotated in
and out of the CPS over a period of 16 months, such that
75 percent of the sample from month to month consists of
the same households and 50 percent from year to year. (See
chapter 1.) Also, sample redesigns and large fluctuations
in the size of the labor force cause changes in the variance
of the estimates. These two features of the CPS—an overlapping sample design and changes in reliability—induce
strong positive autocorrelation and heteroskedasticity in the
standard errors. These characteristics can seriously contaminate estimates of the true labor force if the sampling error is
ignored in the estimation process. For this reason, it is important to specify a model of the standard error process and combine it with the model of the signal, to estimate the unobserved components of the CPS. The standard error model
is specified as:

*e*_{t} = *γ*_{t}e*_{t} ,

with *e*_{t} reflecting the autocovariance structure, assumed to
follow an autoregressive moving average (ARMA) process,
and *γ*_{t}
representing a changing variance over time. The parameters of the ARMA model are derived from sampling
error autocorrelations developed independently of the timeseries model from design-based information. The CPS error
variances are estimated using the method of generalized variance functions (Zimmerman and Robison, 1996).

The unknown hyperparameters of the signal are estimated by maximum likelihood, using the Kalman filter algorithm. Given these estimated hyperparameters, the Kalman
filter is used to decompose the observed CPS into its signal
and noise components. This algorithm efficiently updates
the model estimates as new data become available each
month. For the latest month, the Kalman filter calculates
estimates on the basis of all available data but does not revise estimates for the previous months with the latest data.
Previous estimates are updated by a Kalman filter "smoother," which revises a given period's estimate by using all
data available, both prior and subsequent to the month. This
smoothing procedure is performed only at the end of each
year.

*Real-time Benchmarking*. LAUS model estimates are adjusted by means of a process of real-time benchmarking, whereby each month state model estimates are "ratio adjusted" to
the CPS national estimates of employment and unemployment, such that the not seasonally adjusted estimates for all
states and the District of Columbia sum to the national levels.
By forcing state model estimates to add to current monthly
national CPS estimates, real-time protection is provided for
the models, because the benchmarked estimators will reflect
this change as it occurs, whereas the original model estimators would be slower to adapt. Another benefit of benchmarking is to ensure comparability between estimates at different
levels of geography.

Real-time benchmarking occurs in a two-step process.
First, employment and unemployment estimates for Census divisions are created from CPS-only univariate signal-plus-noise models that are controlled to the national CPS
estimates. Then, state model estimates are controlled to the
appropriate division estimates.

*Smoothed Seasonal Adjustment Process*. In 2010, a smoothed
seasonal adjustment (SSA) process was introduced to reduce
the number of spurious fluctuations in the seasonally adjusted estimates due primarily to noise introduced in real-time
benchmarking. State labor force estimates are smoothed with
the use of the Henderson trend filter, which uses weighted
moving averages to suppress irregular fluctuations in the
seasonally adjusted series, leaving the trend more visible.
Two-sided, symmetric moving averages (up to 13 months in length) are used to smooth the historical series, while a onesided, asymmetric (7-month) average is used in real time.

#### Estimates for substate areas

**Modeled substate areas**

Labor force estimates for the Los Angeles-Long Beach-Glendale, CA, Metropolitan Division (formerly the Los
Angeles-Long Beach metropolitan area) and New York
City have been developed since 1978, first as direct-use
areas and then using bivariate signal-plus-noise models.
In 2005, signal-plus-noise models for five additional
substate areas and their respective state balances were
introduced. The areas are the Chicago-Joliet-Naperville, IL,
Metropolitan Division; the Cleveland-Elyria-Mentor, OH,
Metropolitan Statistical Area; the Detroit-Warren-Livonia,
MI, Metropolitan Statistical Area; the Miami-Miami
Beach-Kendall, FL, Metropolitan Division; and the Seattle-Bellevue-Everett, WA, Metropolitan Division. (Model-based estimation also was initiated for the New Orleans-Metairie-Kenner, LA, Metropolitan Statistical Area but was
discontinued following Hurricane Katrina, due to the storm's
impact on the CPS sample and data collection.) These area
models are based on the classical decomposition of a time
series into trend, seasonal, and irregular components. A
component to identify and remove the CPS sampling error
also is included. Area models, like the Census division
models, are univariate in design in that only the historical
relationship of the CPS inputs is considered—UI claims
counts and CES estimates are not used in the estimation
process. Area and balance-of-state models are controlled
directly to the state benchmarked model totals, which are
themselves controlled to the national CPS via the Census
division models, as described above.

**The handbook method**

Until 1973, the handbook method was the only means used
to develop state and local area labor force and unemployment estimates. With the exception of the seven substate area
models discussed, the handbook method continues to be used
for substate estimation. This method is an effort to use available information to create employment and unemployment
estimates for an LMA that are comparable to what would be
produced by a representative sample of households in that
area, without the expense of conducting a large labor force
survey like the CPS. The method presents a series of estimating building blocks, for which categories of employed and
unemployed workers are estimated and then summed.

*Employment*. The total employment estimate for a particular
area is based on data from several sources. The main sources
are the CES survey, the state-designed monthly survey of establishments, and the Quarterly Census of Employment and
Wages (QCEW), a universe count of employment covered by
the UI system. (See chapter 5.) These sources are designed to
produce estimates of the total number of employees on payrolls in nonfarm industries for a particular area.

Because employment estimates from these sources are
based on the location of the establishment, these "place-of-work" estimates must be adjusted to reflect the place-of-residence concept used in the CPS survey of households.
Resident employment includes workers living and working
in the same area and also those who work in other areas within commuting distance. Estimates of resident employment
should, therefore, reflect employment changes in those related commutation areas as well. In 2005, LAUS introduced
dynamic residency ratios (DRRs) to provide this adjustment.
Multiple residency adjustment ratios were produced, using
Census 2000 county-to-county worker commuting data and
March/April 2000 total nonfarm job estimates. Separate residency adjustment ratios were developed for each estimating area and up to four additional labor market areas into
which at least 100 residents commuted to work. Ratios for
each of the commuting areas are multiplied by their respective monthly nonfarm jobs estimates to produce estimates of
estimating area residents who work in each of the commuting areas. Separate commuting area estimates are summed to
create a total of the resident nonfarm wage-and-salary employed for the estimating area. This adjustment also accounts
for multiple jobholding and unpaid absences in the payroll
employment estimates.

Estimates for employment not represented in the establishment series—agricultural workers and nonfarm self-employed, unpaid family, and private household workers—are
derived by extrapolation from the decennial census. These
components, plus the wage and salary component, represent
total handbook employment. To develop estimates for employment not covered by the establishment series, census
counts are used as base-period estimates to which change factors are then applied. Estimates of nonfarm self-employed,
unpaid family, and private household workers are developed
from change factors based on monthly CES and CPS data.
Agricultural employment estimates are developed from
change factors based on monthly and annual CPS data.

*Unemployment*. The estimate of unemployment is an aggregate of the estimates for each of the two building-block categories: those covered by the UI system and those outside its
scope ("noncovered"). The covered category consists of (a)
those who are currently receiving UI benefits and (b) those
who have exhausted their benefits. The noncovered category
consists of those who are ineligible to receive UI benefits.

A count of the covered unemployed who collected UI benefits during the reference week (the week of the 12th) and
also had no earnings due to employment is obtained directly
from state, federal, and railroad unemployment programs.
Estimates of unemployed persons who have exhausted their
benefits ("exhaustees") are based upon the number of claimants who received their final payments in the week before the
reference week, plus an estimate of exhaustees from previous
periods who are still unemployed. This calculation involves
estimating the percentage of long-term unemployed who
continue to remain unemployed each week and applying that
percentage to the exhaustee pool.

Noncovered unemployed are those persons who are not in the
scope of the UI system. Many of the unemployed were not
employed immediately before their current spell of unemployment and, thus, did not meet the wage and employment
experience requirements to qualify for UI compensation.
Because UI compensation is not a criterion for determining
unemployment status in the CPS, these individuals, known
as entrants to the labor force, are counted as unemployed and
included in LAUS estimates.

Unemployed entrants are divided into two groups: new
entrants and reentrants. New entrants are individuals who
have entered the labor force for the first time. Reentrants are
individuals who have reentered the labor force after a period of neither employment nor unemployment. Both new
entrants and reentrants are estimated from state-level CPS
data on a 5-year weighted-average for each month. Then,
the averages for each state are controlled to the corresponding national total for the current month. Next, these adjusted
statewide entrant totals are distributed to each LMA, using
population shares based on the latest annual July 1 population estimates from the Census Bureau. An LMA new entrant
estimate is calculated as its share of the state population ages
16–19, multiplied by the total statewide new entrant estimate.
Reentrants are estimated by a similar procedure, using each
LMA's share of the state population ages 20 years and older
and the statewide reentrant total.

*Substate adjustment for consistency and additivity*. Each
month, handbook estimates are prepared for LMAs that exhaust each state geographically. To obtain an estimate for a
given area, a "handbook share" is computed for that area; this
is defined as the ratio of the area's handbook estimate to the
sum of the handbook estimates for all LMAs in the state. The
area's handbook share is then multiplied by the current statewide modeled estimate to produce the final adjusted LMA
estimate of employment:

*E*_{a(t)} = *E*_{s(t) *} *HB*_{a(t)} / *ΣHB*_{a(t)} ,

where:

*E* = total employment,

*HB* = handbook employment,

*a* = area,

*s* = state, and

*t* = time.

A comparable adjustment is performed for unemployment.

#### Estimates for parts of LMAs

Current labor force estimates at the sub-LMA level are required by several federal allocation programs. However, the
handbook method was not designed for these small areas,
because the data required to compute independent handbook estimates generally are not available. Based on data availability, two alternative methods are used to disaggregate the
LMA estimates to the subarea level.

The population-claims method is the standard technique
for unemployment. If residence-based UI claims data are
available for the subareas within the labor market area, the
ratio of the subarea to the total number of claims within the
LMA is used to disaggregate the estimate of covered unemployed to the subarea level. The estimate of unemployed
entrants is allocated on the basis of the latest annual distribution of adult and teenage population groups. When the
population-claims method is used for unemployment, employment is disaggregated with the use of current population
distributions prepared by the Census Bureau and weighted by
each area's decennial census relative share of employment to
population. This preferred combination of techniques is used
to derive estimates for all counties in multicounty LMAs and
for cities in nearly all states.

If the necessary UI claims data are not available at the subarea level, the census-share method is used. This method uses
each subarea's decennial census share of county employment
and unemployment, respectively, to disaggregate the monthly
subarea (that is, city) estimates of employment and unemployment.

#### Smoothed seasonal adjustment of Metropolitan Statistical Areas

Employment and unemployment estimates for metropolitan
areas and metropolitan divisions are seasonally adjusted and
smoothed each month, using one of two methods. One option is a non-model-based X-11 Auto-Regressive Integrated
Moving Average (ARIMA) method. The preferred option is
a model-based approach known as Signal Extraction in ARIMA Time Series (SEATS). The approach that yields the best
fit is used for each employment and unemployment time series. Once labor force estimates are seasonally adjusted, they
are smoothed, using the same Henderson trend filter (H13)
used in state estimation. These data were made available on
the BLS website in mid-2011 and will be integrated into news
releases and the time-series database at a later date.

### Annual Processing

At the end of each year, LAUS conducts a review of model
performance. States provide information about their economies. Month-to-month movements and observations are examined to determine if they are reflective of economic events
or if any should be considered outliers.

At the beginning of each year, LAUS receives new population controls from the Census Bureau. CPS estimates for
states, Census divisions, and the United States are revised,
using these new estimates of the civilian noninstitutional
population ages 16 and older. Revisions to state model inputs—specifically, CES and UI—also are received. State and
substate models then are reestimated to incorporate changes
in inputs and population controls, using all data in the series.
Revised statewide estimates are controlled to updated Census division models that sum to national totals, all reflecting the
new population controls. (The official U.S. totals generally
are not updated to reflect new population estimates.) Revised
seasonally adjusted estimates are smoothed with the Henderson-13 symmetric filter, incorporating the current month and
6 prior and subsequent months. Toward the end of the series,
where there are fewer than 6 subsequent months of data, an
asymmetric filter is used.

Substate estimates are revised to incorporate any changes
in the inputs, such as revisions in the place-of-work-based
employment estimates, revisions to claims data, and updated
historical relationships. Area handbook estimates then are
revised and readjusted to sum to the revised state estimates
of employment and unemployment. Areas for which the estimates are disaggregated are revised, using updated population estimates.

### Uses and Limitations

Estimates of unemployment and the unemployment rate are
used by federal agencies to determine the eligibility of an
area for benefits under various federal programs. These include: the Workforce Investment Act (WIA), the Temporary
Assistance for Needy Families (TANF) program, the Emergency Food and Shelter Program (EFSP), The Emergency
Food Assistance Program (TEFAP), the Historically Underutilized Business Zones (HUBZone) program, and Labor
Surplus Area (LSA) designations. Under most programs,
unemployment data are used to determine the distribution of
funds to be allocated to each eligible area. In the case of the
HUBZone and LSA designations, data are used in the determination of area eligibility for benefits.

In 2005, improved time series models introduced reliability measures for the seasonally adjusted and not seasonally
adjusted series and for over-the-month changes. In 2008,
reliability measures were implemented for over-the-year
changes. In early 2007, model-based error measures became
available for annual average estimates. Model-based error
measures are available for regions, divisions, and states. (See
Information on Model-Based Error Measures for Regions, Divisions, and States on the BLS website.) Analysis in the
monthly *Regional and State Employment and Unemployment*
news release reflects the use of these error measures.

Estimates not directly derived from sample surveys or
statistical modeling are subject to errors resulting from the
estimation processes used, as well as the limitations of the
data sources used. The error structure associated with these
estimates is complex, and information on the magnitude of
the overall errors is not available.

*Data products*. Data from the LAUS program are made
available to users in a variety of ways. Labor force and
unemployment data are published monthly for Census regions
and divisions, states, and the model-based substate areas
in a news release entitled *Regional and State Employment and Unemployment*. Estimates for metropolitan areas and
divisions are published monthly in a news release entitled
*Metropolitan Area Employment and Unemployment*.

Annual average data are published each year in a news
release entitled *Regional and State Unemployment*, which
typically is issued in late February. This release presents data
on the population, civilian labor force, employed, unemployed,
and unemployment rate for regions, divisions, and states.
Annual average information for states and metropolitan areas
also is published each spring in *Employment and Earnings Online*.

Current and historical data from the LAUS program for
all 7,300 areas also are available online in the Bureau's public database. Users may access the data via the BLS website
(www.bls.gov/data/) or by download (http://download.bls.gov/pub/time.series/la/). Additional information about the LAUS
program, including frequently asked questions, contacts,
and technical references, are online at the LAUS homepage
(www.bls.gov/lau/).

### Technical References

Bell, W.R., and S.C. Hillmer, "The Time Series Approach to
Estimation for Repeated Surveys," *Survey Methodology*, 1990, pp. 195–215.

Denton, F.T. "Adjustment of Monthly or Quarterly Series to
Annual Totals: An Approach Based on Quadratic Minimization," *Journal of the American Statistical Association*, 1971, Volume 66, pp. 99–102.

Harvey, A.C., *Forecasting, Structural Time Series Models,
and the Kalman Filter*, Cambridge: Cambridge University Press.

LAUS Research Report, Bureau of Labor Statistics. Residency Adjustment for Place-of-Work Nonagricultural
Wage and Salary Employment, March 2004, revised
December 2004.

Scott, A.J., T.M.F. Smith, and R.G. Jones, "Analysis of
Repeated Surveys Using Time Series Methods," *Journal of the American Statistical Association*, 1974,
pp. 674–678.

Tiller, Richard B. "Time Series Modeling of Sample Survey
Data from the U.S. Current Population Survey," *Journal of Official Statistics*, 1992, pp. 149–166.

Tiller, Richard B. "Model-Based Labor Force estimates for
Sub-National Areas with large Survey Errors," Office of
Employment and Unemployment Statistics, Bureau of
Labor Statistics, Washington, DC, 2006 (http://www.bls.gov/ore/pdf/st060010.pdf).

U.S. Department of Labor, Bureau of Employment Security.
*Handbook on Estimating Unemployment, Employment
Security Research Methods, Handbook Series* (BLS Reprint No. R-185), 1960.

U.S. Department of Labor, Bureau of Labor Statistics, *Local
Area Unemployment Statistics Program Manual*,
August 2010.

U.S. Department of Labor, Bureau of Labor Statistics. Questions and Answers on the New Method for Developing
State Employment and Unemployment Estimates, January 1989.

Zimmerman, T.S., T.D. Evans, and R.B. Tiller, "State Unemployment Rate Time Series," in Proceedings of the
Survey Research Methods Section, American Statistical
Association, 1994.

Zimmerman and Robison, 1996, BLS internal memorandum,
"Report on Revised State GVF Parameters for CPS
Monthly Employment and Unemployment Estimates in
States, District of Columbia, and Four Substate Areas:
1976-Present."

**Last Modified Date:** January 31, 2013