[Federal Register Volume 61, Number 47 (Friday, March 8, 1996)]
[Notices]
[Pages 9574-9575]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-5549]




[[Page 9573]]

_______________________________________________________________________

Part V





Department of Labor





_______________________________________________________________________



Bureau of Labor Statistics



_______________________________________________________________________



Notice of Decision To Revise Method for Estimation of Monthly Labor 
Force Statistics for Certain Subnational Areas; Notice

  Federal Register / Vol. 61, No. 47 / Friday, March 8, 1996 / 
Notices  

[[Page 9574]]


DEPARTMENT OF LABOR

Bureau of Labor Statistics


Notice of Decision To Revise Method for Estimation of Monthly 
Labor Force Statistics for Certain Subnational Areas

AGENCY: Bureau of Labor Statistics, Labor.

ACTION: Statement of Policy.

-----------------------------------------------------------------------

SUMMARY: The Department of Labor, through the Bureau of Labor 
Statistics (BLS), is responsible for the development and publication of 
local area labor force statistics. This program includes the issuance 
of monthly estimates of the labor force, employment, unemployment, and 
the unemployment rate for each State and labor market area in the 
nation. Beginning with estimates for January 1996, monthly labor force 
statistics for 11 large States (California, Florida, Illinois, 
Massachusetts, Michigan, New Jersey, New York, North Carolina, Ohio, 
Pennsylvania, Texas) and two large areas (New York City, and the Los 
Angeles Metropolitan Area), are being developed according to the time 
series model approach used in the 39 other States and the District of 
Columbia. This action is in response to a reduction in the number of 
households in the Current Population Survey (CPS) undertaken to address 
lower funding levels for BLS and excessive volatility in the monthly 
CPS estimates for these large States and areas. Historically, the CPS 
sample in these States and areas was sufficiently large to meet the BLS 
standard for direct use and the monthly estimates were taken directly 
from the survey. The BLS will publish monthly estimates for these 
subnational areas based on the time series modeling approach starting 
in March 1996.

DATES: These changes were effective on January 7, 1996.

FOR FURTHER INFORMATION CONTACT:
Sharon P. Brown, Chief, Division of Local Area Unemployment Statistics, 
Bureau of Labor Statistics, telephone 202-606-6390.

SUPPLEMENTARY INFORMATION:

Summary of Comments

    The BLS received 5 comments in response to the proposal to revise 
the method of labor force estimation for certain large States and areas 
which was published November 3, 1995 [60 FR 55855]. Two commenters 
expressed support of the proposal; 2 were opposed. The fifth commenter 
expressed support for model-based estimation, but had reservations 
about the characterization of the model approach and reduction in 
sample size.'
    Three commenters expressed concern that detailed demographic 
statistics from the CPS be preserved. The BLS will continue to make CPS 
demographic estimates available, although the variance of the monthly 
estimates will rise, with a negative impact on analytical uses, because 
of the sample size reduction. These monthly State characteristics data 
will not add up to the official labor force totals which are produced 
by the models.
    Detailed characteristics will continue to be published in the 
annual Geographic Profile of Employment and Unemployment publication. 
These estimates will be consistent with annual totals, but the 
reliability of these estimates will be reduced, and a few may no longer 
be of publishable quality.
    Two commenters asked that parallel estimates be prepared for a 
minimum of one year, to explore other options to maintain the sample 
size and to simulate the effect on the estimates and on federal fund 
allocations. The decision to reduce the sample size of the CPS was made 
because of the anticipated lower funding levels for BLS. Other options 
to achieve commensurate savings were not available. Since the modeling 
approach has been used successfully in the 39 smaller States and the 
District of Columbia, simulation in the larger States was not required.
    One commenter noted that the CPS sample size cut and switch to 
model-based estimation appears to run counter to the purposes of the 
major redesign of the CPS implemented in 1994. BLS does not agree, as 
the models will benefit from the improvements in data accuracy and 
definitional changes stemming from the redesign. A similar concern was 
expressed by a commenter who felt that the models were portrayed as a 
fall-back method. The BLS strongly supports the statistical modeling 
modeling methodology. The models are designed to adapt to changes in 
trend and seasonality in the CPS while using historical relationships 
in the data to smooth current estimates and explicitly removing an 
estimate of the CPS noise. The resultant estimates exhibit considerably 
lower volatility as compared to the sample-based estimates.
    A commenter noted that the CPS estimates have a statistical measure 
of reliability, while the models at this time do not. BLS is 
researching the development of monthly reliability measures for the 
modeled estimates.
    The issue of revision of estimates was raised. Under the model 
methodology, State-wide estimates are revised monthly as well as at 
year-end.
    Operational concerns were expressed by two commenters on the delay 
in the release of data. While the BLS will not publish the State labor 
force estimates until 3-4 weeks after the national release, BLS will 
update the estimating system immediately. Therefore, States will be 
able to make estimates as early as the day that the monthly national 
statistics are released, if they so desire.

Additional Information

    The BLS has been responsible for the Local Area Unemployment 
Statistics (LAUS) program since 1972. In 1978, the BLS broadened the 
use of data from the CPS in the LAUS program by extending the annual 
reliability criterion to monthly data. This action was 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, monthly CPS levels were used directly for the 10 
largest States, two sub-States areas, and the respective balance-of-
State areas. The use of annual average CPS data continued for the other 
40 States and the District of Columbia. Ultimately, the budget proposal 
which initiated the direct use of monthly State CPS data was rejected 
as too costly. Based on population ranking, the State of North Carolina 
joined the group of direct-use States in 1985, bringing the group to a 
total size of 11 States. Also in 1985, sample redesign and other 
efficiencies improved the reliability of CPS data at the State level, 
resulting in the criterion on monthly and annual average data of an 8 
percent coefficient of variation on the level of unemployment when the 
unemployment rate is 6 percent.
    Especially in regard to the monthly direct use of State CPS data, 
concern had been expressed as to the volatility of the statistics. In 
the typical direct-use State, a month-to-month change in the 
unemployment rate had to exceed 0.7 percentage point to be considered 
significantly. Often, States experienced consecutive, offsetting large 
movements in the unemployment rate.
    For the other 39 States and the District of Columbia, after 
extensive research and simulation, variable coefficient time series 
models for monthly estimation of State employment and unemployment were 
introduced in 1989. Further improvements was effected with the 
implementation of signal-plus-noise models in 1994. These models rely 
heavily on monthly CPS data, as well as current wage and salary 
employment 

[[Page 9575]]
and unemployment insurance statistics. At the end of each year, the 
monthly model estimates are rebenchmarked so that the annual averages 
for each State match the annual averages derived directly from the CPS.
    Because of budget reductions, the CPS sample is not of sufficient 
size to provide monthly data directly for the 11 large States, New York 
City, and the Los Angeles Metropolitan Area. Monthly estimates will 
continue to be produced, based on the time series modeling method 
currently used for the other States and the District of Columbia. Data 
for the current direct-use States and areas are no longer released by 
the BLS at the same time as the monthly national labor force 
statistics, but are published about four weeks later in the State and 
Metropolitan Area Employment and Unemployment news release. States that 
are able to do so have the option of releasing these data earlier, 
perhaps even simultaneously with the release of national data. Monthly 
data for these States also are subject to end-of-year benchmarking.
    The impact of the CPS sample cut on the national statistics is to 
increase the variability of most national estimates by about 5 percent. 
For example, under the current sample, a month-to-month change of 0.19 
percentage points in the national unemployment rate represents a 
statistically significant change at the 90-percent confidence level; 
the corresponding change under the former design was 0.18 percent.
    Detailed descriptions of the estimating methods are available at 
the above address.

    Signed at Washington, D.C., this 1st day of March, 1996.
Thomas J. Plewes,
Associate Commissioner for Employment and Unemployment Statistics, 
Bureau of Labor Statistics.
[FR Doc. 96-5549 Filed 3-7-96; 8:45 am]
BILLING CODE 4510-24-M