![U.S. flag](https://www.dol.gov/themes/opa_theme/img/flag-favicon-57.png)
An official website of the United States government
With the simultaneous transition to NAICS and the completion of the CES sample redesign in March 2003, new estimation and publication structures for Statewide estimates and all the CES metropolitan statistical areas (MSAs) needed to be established. A minimum guaranteed publication structure, detailed below, was defined for all States and MSAs. The structure consists of "Expanded Supersectors," which break Manufacturing; Trade, Transportation, and Utilities; and Government into further publication detail. For MSAs with fewer than 100,000 total nonfarm payroll jobs, the minimum publication structure consists of only the aggregate levels of total nonfarm, total private, goods-producing, service-providing, private service-providing, and total government.
All other published series must pass a minimum sufficiency test of at least 30 unique unemployment insurance (UI) accounts in its sample, or a minimum universe employment count of 3,000 with at least 50 percent covered by the sample. The series are tested using first quarter employment data from the Quarterly Census of Employment and Wages program (QCEW, or ES-202) from the most recent year available.
Guaranteed industries not passing the minimum sufficiency test are estimated using a small area model. The model can be considered a generalization of the Fay-Herriot model. Like the Fay-Herriot models, the model uses a linear regression to link direct estimates with predictors (in this case, the same-month five-year average of relative employment change in the benchmark data). However, the new generation model loosens many of the assumptions in the Fay-Herriot, particularly the assumption that the variances are fixed and known. More information on the new model is detailed in the paper "Bayesian Nonparametric Joint Model for Point Estimates and Variances" by Julie Gershunskaya and Terrance Savitsky.
In addition to the guaranteed industries, sectors may also be modeled at the Statewide level. (See below.) Approximately 46 percent of State and area CES series are model-based.
(Statewide and MSAs with more than 100,000 total nonfarm payroll jobs)
Note: Mining and Logging may be combined with Construction if there is insufficient sample and/or universe employment to publish Mining alone. This becomes supersector 15-000000, Mining, Logging and Construction.
(MSAs with fewer than 100,000 total nonfarm payroll jobs)
BLS supports models for these series for Statewide estimates if there is insufficient sample for direct sample-based estimation. States are not required to publish at these levels if they are not significant industries within the State.
Last Modified Date: March 11, 2024