Academic journal article Monthly Labor Review

Using Internal CPS Data to Reevaluate Trends in Labor-Earnings Gaps

Academic journal article Monthly Labor Review

Using Internal CPS Data to Reevaluate Trends in Labor-Earnings Gaps

Article excerpt

The Current Population Survey provides data that are used to compare gaps in the labor earnings of women and men, people of different races, and people of different levels of education; this article presents a data series that uses cell means and more accurately measures gaps and trends in earnings than do other publicly available series

The Current Population Survey (CPS) a large, nationally representative ample of households collected each month since 1942 by the U.S. Census Bureau. (1) this article focuses on data from the surveys conducted in March because the March survey includes an extensive income questionnaire. The data that are publicly available from the CPS are the primary tool used to investigate yearly trends in United States average labor earnings and their distribution. However, to protect the confidentiality of its respondents, the Census Bureau topcodes the highest values from each source of income that it collects when it reports the income in the public-use CPS data. Topcoding is the replacement of a datum representing part or all of a person's true income with a lower value. One of the challenges that topcoding presents for those using the public-use data to examine labor-earnings levels and trends is that the topcodes vary over time, which leads to artificial increases or decreases in earnings (when the term "earnings" appears alone in this article, it still refers to "labor earnings") at the top of the earnings distribution as different fractions of the population are subject to topcoding each year. (2) Although the public-use data are used extensively to measure the earnings gaps between men and women and Blacks and Whites, (3) until now little was known about how topcoding affects comparisons of labor earnings across these subsets of the population. (4)

This article finds that gaps between the earnings of men and women, Blacks and Whites, and people of various education levels are all sensitive to topcoding. Ratios of these earnings as well as trends in the gaps and ratios also are sensitive to topcoding. The article arrives at these findings by analyzing 1975-2007 CPS data and comparing the values of gaps and ratios obtained using the public-use CPS data with values found using the internal CPS data.

This article presents an extended cell mean series that will be explained in more detail in a later section. The earnings gaps calculated using the extended cell mean series in conjunction with public-use CPS data are found to closely approximate those obtained with the Census Bureau's internal CPS data. Additionally, this article finds that women, Blacks, and the less-educated are relatively worse off compared with men, Whites, and the more-educated, respectively, than previously reported using the public-use CPS data. Although the trends for all of the aforementioned earnings gaps are sensitive to topcoding, the impact that attempting to correct for topcoding has on trends differs by year. (5)

Calculating earnings gaps

To calculate gaps in earnings between men and women, between Blacks and Whites, and among people of various levels of education, this article examines the annual labor earnings from wages and salaries, self-employment, and farm earnings of full-time, full-year workers in the CPS. (6) Prior to 1987 these "earnings sources" were reported as three separate values. Since then a fourth source-primary labor earnings (regardless of source)-has been added. The earnings sources and their names in the public and internal CPS data files are listed in table A-1 of the appendix. Much of the previous work exploring earnings gaps between men and women, between or among races, and among people of various levels of education focuses solely on wage and salary earnings and excludes self-employment and farm earnings, primarily because of concerns about the accuracy of self-employment earnings in the CPS. However, as Theresa J. Devine demonstrates, earnings gap data are sensitive to the inclusion or exclusion of self-employment earnings since the earnings gap between men and women is larger among full-time self-employed workers than among full-time wage earners. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.