Using March CPS data to analyze labor market transitions

Article type: Research Article

Authors: Stewart, Jay^a

Affiliations: [a] Employment Research and Program Development Staff, US Bureau of Labor Statistics, 2 Massachusetts Avenue, NE, Room 4945, Washington, DC 20212. Tel.: +1 202 691 7376; Fax: +1 202 691 7425; E-mail: Stewart.Jay@bls.gov

Abstract: Previous research has used short, 2-year panels constructed from matched CPS files to examine a number of topics. This paper shows how researchers can examine labor market transitions using March CPS data without matching, demonstrates that separation rates from these panels are comparable to those generated using the more traditional approach of comparing two points in time, and describes the adjustments required to ensure that the series are consistent over time. The main advantage to using unmatched data, compared with matched data, is the absence of attrition bias and an unbroken time series dating back to 1976. I provide two examples of how these short panels can be used to address issues that would normally require longitudinal data.

DOI: 10.3233/JEM-2007-0288

Journal: Journal of Economic and Social Measurement, vol. 32, no. 2-3, pp. 177-197, 2007