Skip to Main Content
Employment Development Department
Employment Development Department

Methodology for Generating Labor Force Data

Labor Market Information Resources and Data: Home | By Customer | By Subject | By Geography | Data Library | Online Services

The Labor Market Information Division (LMID) uses several methods to estimate statistics for civilian labor force, employment, unemployment, and unemployment rates. These methods discussed in this article were developed in cooperation with the U.S. Department of Labor, Bureau of Labor Statistics (BLS). Additional details and background are available from the BLS website.

Definition of Terms:

Back to Top

CA and LA-LB-Glendale MD - Time Series Models

In January 1996, time series models replaced the Current Population Survey (CPS) as the basis for the estimates of labor force data (labor force, employment, unemployment, and the unemployment rate) for California. In January 2005, the LMID revised data back to 1976 using the new time series models. The models cover two areas of the State: the Los Angeles-Long Beach-Glendale Metropolitan Division (MD) and the "Balance of California" (i.e., the rest of California). The results are added together to derive state-level data.

The time series models consist of two models for each area (Los Angeles-Long Beach-Glendale MD and Balance of California):

With these data and estimates of population change, employment, unemployment, and labor force are calculated. The models estimate ratios (employment-to-population and the unemployment rate) rather than the employment and unemployment levels because these ratios are easier to estimate than specific levels.

Back to Top

Unemployment Rate Model

The unemployment rate model uses the relationship between the monthly Unemployment Insurance (UI) claims data and the CPS unemployment rate.

Flexible trend and seasonal components are included to account for movements in the CPS rate that are not reflected in the historical UI claims series.

Back to Top

Employment-to-Population Model

The employment-to-population model uses the relationship between the ratio of the monthly Current Employment Statistics Survey (CES) employment to the population and the ratio of CPS employment to the population.

The model also includes trend and seasonal components to account for movements in the CPS not captured in the CES series. The seasonal component accounts for the seasonality in the CPS not explained by the CES (for example, agricultural employment movement), while the trend component adjusts for long-run systematic differences between the two series (for example, during expansions, the CES grows faster than the CPS).

Under the time series models for the Los Angeles-Long Beach-Glendale MD and the Balance of California, the previous month's estimates are revised. State monthly model estimates are controlled using "real-time" benchmarking to the national monthly labor force estimates from the CPS. This reduces the regular annual revisions at the end of the calendar year to the state unemployment and unemployment series.

Back to Top

Substate Labor Force Data - LAUS Method

The time-series models, discussed earlier, produce state-level data as well as data for the Los Angeles-Long Beach-Glendale MD. Estimates for substate areas, except Los Angeles-Long Beach-Glendale MD, are produced using indirect estimation techniques described below.

In the Local Area Unemployment Statistics (LAUS) program, the LAUS Handbook Employment and Unemployment method is used for producing sub-state employment and unemployment estimates.

Total Nonagricultural wage and salary employment from the CES (adjusted for residency using the 2010 Census)

Total Unemployment Insurance (UI), Unemployment Compensation for Federal Employees (UCFE) and Railroad Retirement Board (RRB) claims less earnings

Back to Top

Sub-County Areas - Census Share Method

The LMID derives monthly labor force data for sub-county areas by using the share of county-level employment and unemployment in the area at the time of the most current five-year American Community Survey (ACS) estimates, which are updated annually.  The sub-area’s employment and unemployment estimates are then added to determine the total labor force and unemployment rate.

This method assumes that the rates of change in employment and unemployment are exactly the same in each sub-county area as at the county level (the same process is used for unemployment). If this assumption is not true for a specific sub-county area, then the estimates for that area may not be representative of the current economic conditions. Since this assumption is untested, caution should be employed when using these data.

Back to Top

Cautions When Using These Data

Back to Top