Revisions in Hispanic population and labor force data.
Based on information from the 1980 census, independent population estimates for Hispanics were developed for January 1980 up through the present. This, in turn, permitted a revision of the historical data for major Hispanic labor force series for this period. (Data prior to 1980 are not comparable to the revised series.) Monthly seasonally adjusted data for the two independently adjusted Hispanic series--employment and unemployment levels for all Hispanics age 16 and over--have also been revised back to 1980. From these, adjusted labor force, participation rate, employment-population ratio, and unemployment rate series are derived.
In the past, the CPS did not use independent population estimates for Hispanics--the only major population group for which this was the case. Instead, the population estimates were derived from the CPS itself. This yielded estimates that were too low relative to those from the decennial census (because of problems with CPS coverage) and quite unstable over time. Under the revised procedure, CPS sample estimates are "inflated" to the independent estimate of the Hispanic population rat her than being determined by the proportion of Hispanics found in the sample each month.
The independent population estimates were developed using a cohort-component methodology, in which the 1980 census count is updated by adding estimates of Hispanic births and immigrants and subtracting estimates of deaths and emigrants. These procedures integrate data on changes in the Hispanic population from a number of sources. Data on births come from the annual CPS fertility questionnaire and from the National Center for Health Statistics. Death rates are derived from mortality stat istics in California and Texas, States with more than half of the Hispanic population in 1980. Data on immigration and emigration are from the Immigration and Naturalization Service, the Puerto Rican Planning Board, and the Office of Refugee Resettlement.
The new methodology results in sharply higher population estimates and, hence, higher labor force counts, although overall national estimates are not affected. For example, table 1 shows that, on an annual average basis for 1984, the revised Hispanic civilian noninstitutional population levels were almost 1.3 million, or 13 percent higher than the old estimates. Adult men were the group most affected by these changes; their 1984 population estimates rose by more than 18 percent. The levels of various labor force measures (that is, employment, unemployment, and persons not in the labor force) expanded, to a large extent, proportionately. Hence, rates calculated using these levels are not significantly different from those derived with the old methodology. For example, in 1984, only the unemployment rates for teenagers rose by more than a tenth of a percentage point. Revised data for major Hispanic labor force measures for the years 1980-84 are available upon request.
|Printer friendly Cite/link Email Feedback|
|Author:||Rones, Philip L.|
|Publication:||Monthly Labor Review|
|Date:||Mar 1, 1985|
|Previous Article:||New data series on involuntary part-time work.|
|Next Article:||Postal negotiations end in arbitration.|