Is Job Stability in the US Falling? Reconciling Trends in the Current Population Survey and Panel Study of Income Dynamics
David A. Jaeger, Ann Huff Stevens
Abstract
Open-access reader
David A. Jaeger, Ann Huff Stevens
Abstract
Open-access reader
Documenting trends in job stability over the past twenty-five years has become a controversial exercise. The two main sources of information on employer tenure, the Panel Study of Income Dynamics (PSID) and the Current Population Survey (CPS), have generally given different pictures of the degree of job stability in the U.S. economy. This paper examines whether the PSID and CPS yield systematically different results with respect to comparable measures of job stability. We find that there is little evidence in either data set of a trend in the share of employed individuals with one year or less of tenure. Both data sets do show an increase in the fraction of male workers aged 30 and over with tenure less than ten years beginning in the early 1990's. We find that the two data sets provide nearly identical results for the 1980's and 1990's while in the 1970's they give results that are somewhat less comparable. We argue that this is probably the result of changes in the CPS tenure question following the 1981 survey. The effects of this change and the choice of ending year and variable definition in PSID-based studies are the most likely explanations for the disparate findings in the literature.
OpenAlex reports 11 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Documenting trends in job stability over the past twenty-five years has become a controversial exercise. The two main sources of information on employer tenure, the Panel Study of Income Dynamics (PSID) and the Current Population Survey (CPS), have generally given different pictures of the degree of job stability in the U.S. economy. This paper examines whether the PSID and CPS yield systematically different results with respect to comparable measures of job stability. We find that there is little evidence in either data set of a trend in the share of employed individuals with one year or less of tenure. Both data sets do show an increase in the fraction of male workers aged 30 and over with tenure less than ten years beginning in the early 1990's. We find that the two data sets provide nearly identical results for the 1980's and 1990's while in the 1970's they give results that are somewhat less comparable. We argue that this is probably the result of changes in the CPS tenure question following the 1981 survey. The effects of this change and the choice of ending year and variable definition in PSID-based studies are the most likely explanations for the disparate findings in the literature.
Key concepts: Panel Study of Income Dynamics, Current Population Survey, Demographic economics, Falling (accident), Stability (learning theory), Panel data, Economics, Population