FAQs Regarding Recent Changes to Labor Force Estimates
1) What changes are being made to the way unemployment rates are calculated for Texas?
The Labor Market & Career Information Department (LMCI) of the Texas Workforce Commission produces labor force estimates each month using methods mandated by the U.S. Bureau of Labor Statistics (BLS). The BLS is implementing a redesigned method for producing labor force estimates beginning with the January 2005 estimates that are scheduled for release on March 10, 2005. The redesigned method includes an improved time-series regression model as well as real-time benchmarking. Real-time benchmarking to the national Current Population Survey (CPS) will ensure that estimates of employment and unemployment for the states add to the national totals. In addition to reducing end-of-year revisions, it also means that major economic events will be reflected in a more timely manner in the state estimates.
Another change includes the use of more geographically precise unemployment claims data. Under this new system, the LMCI Department is able to better pinpoint the locations of the unemployed. We can now produce more accurate unemployment rates for local areas, such as counties and cities. Though the number of unemployment benefit recipients is not the only input used in producing estimates of the number unemployed in an area, it is an important element.
Taken together, all of the above changes being implemented by BLS will improve estimate responsiveness to changes in the overall economy.
2) Why are these changes being made?
First and foremost, changes are being made to update the labor force estimates to the 2000 Census-based information. Regardless of the redesign method, the Bureau of Labor Statistics (BLS) would still have made changes to incorporate the new Metropolitan Statistical Area (MSA) geographic designations developed by the Office of Management and Budget and used in the 2000 decennial census.
In addition, BLS is implementing the redesign method in order to address a number of significant and long-standing issues that have been identified with the previous method of model estimation and annual benchmarking. The redesigned method includes an improved time-series regression model and introduces real-time benchmarking. Real-time benchmarking to the national Current Population Survey (CPS) will ensure that estimates of employment and unemployment for the states add to the national totals. In addition to reducing end-of-year revisions, it also means that major economic events will be reflected in a more timely manner in the state estimates.
3) Were the previous estimates produced by LMCI wrong?
The labor force estimates previously produced by the LMCI Department were not wrong; they were the most accurate figures that could be produced using the methods and inputs mandated by the Bureau of Labor Statistics at the time. The previous modeling and benchmarking approach reflected state-of-the-art methodology and operations in 1994. Those estimates are still considered valid, though not directly comparable to estimates for the period January 2005 and later.
4) Who decides what methods are used to produce estimates of an area's labor force?
The U.S. Bureau of Labor Statistics (BLS) develops the methods that are required for use by state partners, such as the Texas Workforce Commission. BLS funds the Local Area Unemployment Statistics (LAUS) program in each state to produce monthly labor force estimates for that state. In this way, all states use the same BLS-provided methodology, ensuring that comparable estimates are produced nationwide.
5) Will these changes cause a noticeable difference in unemployment rates, i.e., will they be higher/lower than before?
These changes may produce unemployment rates that are higher or lower than in previous periods for sub-state levels, such as MSAs, counties, or cities.
For Texas, the labor force estimates will not have a break in the time series. However, that is not the case for sub-state estimates.
For sub-state estimates, labor force estimates produced using the new methodology will not be comparable to estimates produced under the previous system. This means that data produced prior to January 2005 will not be comparable to estimates for January 2005 and later. There will be a break in the labor force data series between December 2004 and January 2005.
7) Will data still be produced under the old MSA definitions?
In addition to changes in methodology and inputs, the BLS will also incorporate the new Metropolitan Statistical Area (MSA) geographic designations developed by the Office of Management and Budget and used in the 2000 decennial census. Beginning with the release of January 2005 estimates, data will only be produced under the new MSA definitions. BLS and the LMCI Department will recalculate historical estimates under the new MSA definitions. Data for current and historical periods will no longer be available under the old MSA definitions. Previously, Texas had 27 MSAs. Under the new definitions, the number will be reduced to 25. In addition to geographic changes, some MSAs will also undergo name changes.
Link to table comparing old and new MSA definitions
Link to article regarding MSA changes
Link to Office of Management and Budget bulletin
8) How far back will data produced using this new methodology be available? Will LMCI recalculate labor force estimates "backwards" so that a time series will be available for comparison purposes?
At the statewide level, comparable labor force estimates will be available back to 1976, including seasonally-adjusted and actual (unadjusted) figures. There will not be a break in the statewide data series.
The Bureau of Labor Statistics (BLS) will produce recalculated labor force estimates for the period 1990-2004 for sub-state areas. This data will be available on the BLS website beginning March 10, 2005.
The 1990-2004 data produced by BLS will be based on the original estimates produced for that time series; however, it will reflect the 2000 Census-based geography. In addition, the data will contain model inputs to provide a smoother transition between 1999 and 2000.
Though recalculated, the BLS sub-state data will NOT be comparable to estimates for January 2005 and later. The LMCI Department will release benchmark revised estimates for 2000, 2001, 2002, 2003, and 2004. However, these revised years will not be available all at once. The tentative schedule for releasing labor force benchmark revisions are:
Once the BLS has released its recalculated data and the LMCI Department has released its benchmark revisions for 2000 through 2004, the timeline of available data will be as follows:
Note: Once they are released, the revised estimates produced by the LMCI Department for 2000 through 2004 should be used instead of the BLS data for these years. BLS will replace their recalculated estimates with LMCI's benchmark revised estimates for the 2000-2004 period at a later date. Though revised, LMCI's 2000-2004 estimates will still not be comparable to estimates for January 2005 and later.
9) What is the release schedule for LMCI's revised labor force data?
The tentative schedule for releasing labor force benchmark revisions are:
These dates are subject to change.
10) What changes, in particular, are being made to the way unemployment rates are calculated for cities?
With improvements in geocoding capability, the LMCI Department will be adopting a new BLS-approved method for producing city-level unemployment rates. In the past, the LMCI Department used a census-share methodology to derive city-level labor force estimates. This involved using census ratios (city employment/unemployment as a ratio of county employment/unemployment) from the decennial census and multiplying these ratios by more current county-level employment/unemployment estimates. This method had a significant weakness: It held the ratio of city-to-county employment/unemployment constant from the last decennial census until the release of the next decennial census data, which was up to 15 years later. We know that a number of suburban counties outside metropolitan areas saw significant population growth in the 1990s, therefore the 1990 ratios likely no longer accurately described the local labor market picture. In the past, this was the only method approved by BLS for states to produce city-level labor force estimates.
Under the new methodology, LMCI will use the 2000 Census-based Employment-to-Population Ratio - multiplying this ratio by current population estimates (annual estimates available from the U.S. Census Bureau). This means that the city-level unemployment rates will now be based upon more current (annual) population estimates, and therefore will be much more timely, giving a better picture of the labor force changes occurring in those areas experiencing rapid population growth.
Under the new, BLS-mandated methodology, in order for city-level labor force estimates to be produced, a city must meet three criteria:
While the updated methodology will produce more reliable estimates of an area's labor force, these requirements also mean that the number of cities for which monthly estimates are produced will be dramatically reduced. Under the previous methodology, LMCI produced estimates for 359 cities in Texas. Under the new system, the number of BLS-approved cities will fall to 97. These changes may also produce unemployment rates for areas that are higher or lower than in previous periods, simply due to the redesigned methodology.
11) Why is LMCI no longer producing unemployment rates for my city? How many cities will LMCI be producing labor force estimates for?
Under the new, BLS-mandated methodology, in order for city-level labor force estimates to be produced, a city must:
While the updated methodology will produce more reliable estimates of an area's labor force, these requirements also mean that the number of cities for which monthly estimates are produced will be dramatically reduced. Under the previous methodology, LMCI produced estimates for 359 cities in Texas. Under the new system, the number of BLS-approved cities will fall to 97.
Link to list of BLS-approved cities
12) Where can I get more detailed information about the recent changes in the way labor force estimates are produced?
For more detailed information regarding these changes, please see the Bureau of Labor Statistics' webpage Questions and Answers on the Local Area Unemployment Statistics (LAUS) Program Redesign for 2005.
13) Who can I contact if I have a question that is not answered here?
Please feel free to contact LMCI Department staff with any questions you have regarding these changes.
Labor Market & Career Information Department
Texas Workforce Commission
101 E. 15th St, Suite 103A2
Austin, TX 78778-0001
Phone: 1-866-938-4444 (toll free in Texas)
(512) 491-4922
Fax: (512) 491-4904
Email: