E9-27930. Solicitation of Comments on a Proposal To Revise Method for Estimation of Monthly Labor Force Statistics for Certain Subnational Areas

AGENCY:

Bureau of Labor Statistics, Labor.

ACTION:

Notice of solicitation of comments.

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. A hierarchy of estimation methods is used to produce the 7,300 estimates covered by the Local Area Unemployment Statistics (LAUS) program (http://thomas.loc.gov/​bss/​d111/​d111laws.html), based on the availability and quality of data from the Current Population Survey (CPS). The strongest estimating method—signal-plus-noise models with real-time benchmarking for current estimation and historical benchmarking—is employed for all States and the District of Columbia, the Los Angeles-Long Beach-Glendale, CA metropolitan division, New York City, and the respective balances of New York and California. Models are also employed for five additional substate areas and their State balances. The areas are: the Chicago-Naperville-Joliet, IL metropolitan division; the Cleveland-Elyria-Mentor, OH metropolitan area; the Detroit-Warren-Livonia, MI metropolitan area; the Miami-Miami Beach-Kendall, FL metropolitan division; and the Seattle-Bellevue-Everett, WA metropolitan division.

As part of a program of continuing improvements in LAUS methodology, BLS is proposing the implementation of smoothed-seasonally-adjusted series for current and historical estimates. This approach is an innovative alternative to an annual historical benchmark for seasonally-adjusted State estimates that will address longstanding issues related to end-of-year revision, and also will enhance the analytical capability of the estimates.

BLS proposes to implement the revised methodology beginning with January 2010 current estimates, with historical estimates revised to 1976 for States, the District of Columbia, Los Angeles, New York City, and the respective balances of California and New York. The five other substate model estimates will be revised back to 1983.

DATES:

Submit written comments on or before December 21, 2009.

ADDRESSES:

Send comments to Sharon Brown, Division Chief, Division of Local Area Unemployment Statistics, Bureau of Labor Statistics, Room 4675, 2 Massachusetts Avenue, NE., Washington, DC 20212, by FAX at 202-691-6459, or by e-mail at [e-mail].

FOR FURTHER INFORMATION CONTACT:

Sharon Brown, Division Chief, Division of Local Area Unemployment Statistics, Bureau of Labor Statistics, Room 4675, 2 Massachusetts Avenue, NE., Washington DC 20212, by telephone at 202-691-6390, or by e-mail at [e-mail].

SUPPLEMENTARY INFORMATION:

I. Background

The Department of Labor, through the BLS, has been responsible for the development and publication of local area labor force statistics since 1972. In 1978, the BLS broadened the use data from the CPS in the LAUS program by extending the annual reliability criterion to monthly data. Monthly CPS levels were used directly for the 10 largest States, two substate areas (New York City, Los Angeles), and the respective balances of New York and California. In 1985, the sample redesign and other efficiencies improved the reliability of CPS data at the State level, resulting in the current 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. In addition, North Carolina joined the group of direct-use States. In 1989, variable coefficient time-series models for monthly estimation of State employment and unemployment were introduced for 39 States and the District of Columbia. Further improvement 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 and unemployment insurance statistics. State labor force estimation for the direct-use States was based the time series modeling approach beginning in January 1996.

Improvements introduced with the redesign in January 2005 ensured that State estimates add to the national estimates of employment and unemployment each month, through real-time benchmarking. In doing so, the benchmark changed from annual State-level estimates of employment and unemployment to monthly national estimates of these measures. In this way, economic shocks are reflected in the State estimates on a real-time basis, and end-of-year revisions are significantly smaller.

Historical benchmarking is part of the annual processing activities performed on the models. The first two steps, revision of inputs and model re-estimation, are the same for both the not-seasonally-adjusted (NSA) series and the seasonally-adjusted series. The final step, benchmarking to historical control totals, differs by series. The NSA estimates are benchmarked to monthly Division model controls which have been controlled to monthly national CPS estimates. This ensures that the monthly State NSA estimates sum to the national CPS estimates. The annual average of the NSA estimates is used to control the monthly seasonally-adjusted model estimates. This process preserves the underlying smoothness in the model estimates that would be lost by applying the monthly benchmarking procedure.

However, the current procedure had an unanticipated impact on the historical benchmarking for the seasonally-adjusted estimates during 2008. Unemployment rose steeply in the nation and all States during 2008. The benchmark methodology that required the use of the annual average as the historical control total for the seasonally-adjusted estimates meant that unemployment rates were adjusted downward during the latter months of 2008. This impacted comparisons with January 2009 unemployment estimates that continued to reflect the steep economic decline. In addition to issues with historical benchmarking, the monthly real-time benchmarking procedure introduces volatility into the current seasonally-adjusted estimates, producing estimates with spurious turning points that are difficult to explain to data users.

II. Current Action

To address these serious issues, the BLS proposes modifying the procedures for the seasonally-adjusted estimates and implementing a smoothing methodology for both current and historical seasonally-adjusted series. Smoothing the current series will reduce the number of spurious turning points in the estimates. For historical estimates, the first two steps in annual processing: revising model inputs and re-estimating the series, are unchanged. The last step, benchmarking to control totals, will be revised for the seasonally-adjusted estimates. The use of the annual average of the NSA series as the control total will be dropped. Instead, as in current monthly estimation, the historical seasonally-adjusted series will be adjusted by the same pro-rata factor used in adjusting the NSA estimates to the national control totals. Since the pro-rata factors fluctuate from month-to-month, this procedure will introduce additional variability into the historical series, which could dominate the monthly change in the benchmarked series. Smoothing the series following the application of the pro-rata adjustment will reduce the volatility added. The smoother selected is the Henderson Trend Filter (H13).

Detailed descriptions of the current and proposed approaches are available from the office listed above.

III. Desired Focus of Comments

Comments and recommendations are requested from the public on the use of the Henderson Trend Filter (H13) to smooth the LAUS current and historical seasonally adjusted estimates.

Signed at Washington, DC, this 17th day of November 2009.

Kimberley Hill,

Acting Chief, Division of Management Systems, Bureau of Labor Statistics.