Job Turnover Heterogeneity and Person-Job-Specific Time-Series Wages

This paper introduces and tests for the endogeneity of job tenure and number of jobs in the wage equation using panel data uniquely suited for that purpose – the National Longitudinal Survey of Youth (NLSY). The model is multi-level in that there is wage dynamics – random intercept, random growth path and autocorrelated transitory variation – at the person-career level and at the job-specific level. The NLSY wage data is unique in that wage values are associated with a specific job (employer) and thus can be used to form a (monthly) time series of job-specific wages. The resulting job-specific time series of wages includes a variable number of unequally spaced points in time and has more missing values than complete data points from beginning to end. The multiple job-specific time series for a person may overlap in time so that the person-specific time series may include multiple wages at any given survey date. The data is also unique in that the begin and end dates of all jobs are available over the person’s career so that precise measures of number of jobs and of months of work experience and tenure on the job may be constructed as of the date of each reported wage value. This data also allows the modeling of job turnover and job duration in continuous time jointly with the wage time series for that job. The model is designed to test the exogeneity of number of jobs held and of job tenure in the wage equation.