UC San Diego

This dissertation develops new data and methods for properly measuring U.S. labor market dynamics using large, nationally-representative household surveys. These data are used to assess potential biases arising from time aggregation and from geographic mobility. Time aggregation is estimated using weekly labor force information from the Survey of Income and Program Participation (SIPP). The degree of time aggregation is large: gross flows estimated from monthly data understate the true number of transitions by 20 percent on average. However, time aggregation creates no meaningful cyclical bias in measured gross flows or hazard rates. Separation hazard rates calculated from the SIPP and the Current Population Survey (CPS) are strongly countercyclical and remain so after adjusting for time aggregation. Using a new database that captures all longitudinal information in the CPS individuals who move can be identified. Comparing the behavior of the entire CPS sample with the subset known not to have moved provides a bound to the bias from geographic mobility. The cyclical bias from geographic mobility is small. At business cycle frequencies, the difference between the separation hazard rate calculated from the entire CPS sample and from a subset that are known not to have moved never exceeds 4 percent. There is little effect of mobility on the job finding hazard rate. The weekly SIPP data identify direct employment-to- employment (EE) transitions. Abstracting from labor force participation, EE transitions account for one-half of all separations from employment. Similar estimates using the CPS are twice as large however the CPS overstates EE transitions because of time aggregation. Separations to a new job are strongly procyclical while separations to unemployment are strongly countercyclical. The combination yields a nearly acyclical total separation rate. The weekly job finding rate is strongly procyclical