IZA DP No. 10245: Rotation Group Bias in Measures of Multiple Job Holding
published in: Economics Letters, 2016, 147 (October), 160-163.
Reported multiple job holding rates in the U.S. are found to be substantially higher among workers in their first month in the CPS sample (the first rotation group), with rates declining in subsequent rotation groups. True rates should not differ across rotation groups. Using 22 years of CPS data, multiple job holding rates based solely on the first rotation group were 27.5 percent higher than official rates based on all rotation groups. Rotation group bias worsened over time and could account for as much as one-quarter of the measured decline in multiple job holding.