Labor Market Transitions and Geographic Mobility: Implications for Worklife Studies

Standard worklife estimates use CPS matched samples to estimate transition probabilities. A potential problem with this methodology is that the basic unit in the CPS survey is a particular address, not a respondent or household. Individuals or households who move are excluded from the sample. If there is a systematic relationship between moving and employment transitions, use of the CPS sample will bias estimates of transition probabilities and expected worklives.

The SIPP survey has the advantage of following individuals, even if they change addresses. Using the SIPP data we estimate the differing transition probabilities faced by movers and non-movers and evaluate the potential bias in the CPS samples. We find, depending on age and sex, interesting and statistically significant differences in transition probabilities exhibited by movers. However, the net effect on estimates of worklives turns out to be relatively small.