Historical Comparability of Illinois Statewide and Sub-State Local Area Unemployment Statistics Data

The federal-state Local Area Unemployment Statistics (LAUS) program has operated under the direction of the U.S. Bureau of Labor Statistics (BLS) since the 1970s. During this time, the methods and procedures for developing statewide and sub-state unemployment statistics have been updated, resulting in occasional discontinuities or breaks in the historical monthly and annual time series. What follows is a description of the most significant changes introduced to statewide and sub-state estimation. This information should help data users interpret what changes to the statewide and sub-state historical series were statistical as opposed to economic in nature.

Monthly statewide estimates are developed from statistical models designed by the U.S. BLS that rely heavily on data from the monthly Current Population Survey (CPS) and to a lesser extent monthly total non-farm jobs and Unemployment Insurance (UI) claims inputs. The official Illinois statewide monthly, annual time series begins in 1976 and is methodologically consistent over time with two exceptions. In 1994, a redesign of the CPS was introduced that included changes to labor force status questions and data collection methods. These changes had an indirect effect on statewide, model-based estimates. Also, each year CPS estimates are adjusted or controlled to the latest working-age, non-institutional population estimates from the U.S. Census Bureau. New population controls are applied to monthly CPS employed, unemployed estimates for the previous five years and impact statewide model-based employed, unemployed estimates. There can also be relatively large breaks in statewide employed, unemployed estimates at the beginning of the decade when population estimates are adjusted to decennial census benchmarks.

Chicago Metropolitan Division estimates are developed from statistical models designed by the U.S. BLS that rely on historical and monthly data from the monthly CPS. The official Chicago-Naperville-Joliet Metropolitan Division monthly, annual labor force, employed, unemployed and unemployment rate series begins in 1983 and is methodologically consistent over time with the two exceptions noted above for Illinois, namely the CPS redesign in 1994 and updates to CPS population controls.

We have also provided data users with unofficial monthly and annual Chicago-Naperville-Joliet Metropolitan Division estimates for years 1976-1982. The Chicago area estimates for 1976-1982 were developed by summing county estimates to the Metropolitan Division level. The 1976-1982 estimates for the eight-counties in the Chicago Metropolitan Division were produced using county disaggregation methods. See description of employment and unemployment disaggregation methods at Understanding Labor Force Statistics. Thus, the 1976-1982 estimates are not strictly comparable to the official, model-based series that begins in 1983.

Estimates for Metropolitan Areas (other than the Chicago-Naperville-Joliet Metropolitan Division) are developed using the Handbook method and then forced to sum to independent, model-based estimates for the portion of Illinois that falls outside the Chicago-Naperville-Joliet Metropolitan Division. For more information on Handbook method and other aspects of metropolitan area estimation visit Understanding Labor Force Statistics.

The official, BLS-approved metropolitan area time series begins in 1990, but we have provided data users with unofficial metropolitan area estimates for years 1976-1989. There have been breaks in the metropolitan area time series as a result of changes to metropolitan area geographic boundaries, updates to population and employment benchmarks following the decennial census and changes in Handbook methodology. Metropolitan Area estimates for years 2000-forward are methodologically consistent. However, estimates for years prior to 2000 are consistent in terms of geographic boundaries but not in terms of methodology.

In 2000, changes to the Handbook method were made for estimating the resident employed and labor force entrant unemployed. These methodological changes resulted in breaks in metropolitan area employed and unemployed estimates between years 1999 and 2000.

Metropolitan area estimates for years prior to 2000 were developed by summing county estimates to 2000-based metropolitan area levels. Some of the county estimates were developed using the Handbook method while other county estimates were produced using disaggregation methods. Also, estimates for years prior to 2000 relied on updates to employment and population benchmarks from the 1970, 1980 or 1990 decennial censuses. This mixture of county estimation methods and updates to decennial census benchmarks resulted in time series breaks in employed and unemployed estimates not only between 1999 and 2000 but also 1989 and 1990 and 1982 and 1983. There was also a break in employed estimates between 1984 and 1985 because of changes to county disaggregation methods (described below).

County estimates are developed using the Handbook or disaggregation methods. The official, BLS-approved time series for counties begins in 1990 but we have provided data users with unofficial county estimates for years 1974-1989. County estimates for years 2000 forward are methodologically consistent. There is a time series break between 1999 and 2000 due to changes in decennial census benchmarks, changes to Handbook methodology as well as conversion from Handbook to disaggregation methods or vice-versa. Also, beginning in 2000, the disaggregation method for county unemployed estimates was updated. The labor force entrant portion of unemployed is now disaggregated from statewide estimates of youth and adult age-group unemployed using annual age group population estimates. Prior to 2000, labor force entrant unemployed were disaggregated to counties using the proportion of age group population estimates from the most recent decennial census. Beginning in 2002, a new method was introduced which improved the accuracy of county residency assignment on UI claims used in county unemployed disaggregation. Updates to decennial census benchmarks and changes from Handbook to disaggregation methods (or vice-versa) resulted in a time series break between years 1989-1990. There is a break in series between 1984 and 1985 due to the introduction of the population-claims share disaggregation method. Population-claims share is described in more detail under employment and unemployment disaggregation at the following page: Understanding Labor Force Statistics. Prior to 1985, county employed and unemployed estimates were disaggregated using fixed employed and unemployed ratios developed using data from the most recent decennial census - also known as census share.

City estimates are developed for communities with at least 25,000 residents using the disaggregation method. The official, BLS-approved time series for cities begins in 1990 but we have provided data users with unofficial monthly and annual estimates for most cities back to 1974. City estimates for years 2000 forward are methodologically consistent. There is a time series break between 1999 and 2000 due to changes in decennial census benchmarks and the disaggregation method for city unemployed estimates. Beginning in 2000, the labor force entrant portion of unemployed was disaggregated from county estimates of youth and adult age-group unemployed using annual city age group population estimates. Prior to 2000, labor force entrant unemployed were disaggregated from counties to cities using the proportion of age group population estimates from the most recent decennial census. Beginning in 2002, a new method was introduced which improved the accuracy of both city and county residency assignment on UI claims used in city to county unemployed disaggregation. Updates to decennial census benchmarks resulted in a time series break between years 1989-1990. There was a break in series between 1984 and 1985 due to the introduction of the population-claims share disaggregation method (Understanding Labor Force Statistics for more information on disaggregation methods). Prior to 1985, city employed and unemployed estimates were disaggregated from counties using fixed employed and unemployed ratios developed using data from the most recent decennial census - also known as census share.

• Introduction of Current Population Survey (CPS) redesign in 1994 including changes to labor force status questions and data collection methods
• Breaks in time series as a result of new population controls being introduced for just part of historical series (usually prior five years) and shifts to decennial census population benchmarks
• Chicago-Naperville-Joliet Metropolitan Division
• Introduction of Current Population Survey (CPS) redesign in 1994 including changes to labor force status questions and data collection methods
• Breaks in time series as a result of new population controls being introduced for just part of historical series (usually prior five years only) and transition to decennial census population benchmarks
• Beginning with 1983 estimates, Chicago Metropolitan Division estimates produced with BLS statistical model that uses historical and current monthly CPS data as inputs.

• Updates to Handbook employed, unemployed estimation introduced beginning with the year 2000
• Breaks between 1999-2000, 1989-1990 and 1982-1983 because of differences between methods for producing estimates for counties within metropolitan areas (i.e. Handbook and disaggregation methods) after metropolitan area geographic boundaries updated following decennial census
• Updates to decennial census benchmarks and inputs, such as population totals, employed estimates and residency adjustment factors in 1983, 1990 and 2000

• Breaks between 1999-2000, 1989-1990 due to updates to decennial census employed and population inputs used in county disaggregation method
• Break between 1999-2000 because of methodological changes to county unemployed disaggregation
• Breaks between years 1999-2000, 1989-1990 and 1983-1983 for counties that had methods switched from Handbook to disaggregation because of updates to metropolitan area boundaries following decennial census
• Improved method introduced for assigning county of residence to Unemployment Insurance claims used in county unemployed disaggregation beginning with estimates for 2002
• Break between 1984-1985 because of introduction of population-claims share method for disaggregating county estimates

• Breaks between 1999-2000, 1989-1990 due to updates to decennial census employed and population inputs used in city-to-county disaggregation method
• Break between 1999-2000 because of methodological changes to city unemployed disaggregation
• Improved method was introduced for assigning city of residence to Unemployment Insurance claims used in city-to-county unemployed disaggregation beginning with estimates for 2002
• Break between 1984-1985 because of introduction of population-claims share method for disaggregating city estimates​