This paper has analyzed the effects of job mobility on earnings both at young and at older ages. The discussion In Part II developed an earnings function which took Into account: the discontinuity of earnings across jobs, the decline of human capital Investment within the job and over the life cycle, and the effects of mobility on the slope of the earnings profile. It was shown that careful attention to the functional form of the equation led to an understanding of why the coefficient of the current segment Is usually larger than the coefficient of the previous segments. Using the expanded equation on the NLS surveys led to several major empirical findings:

1. Mobile Individuals at all ages Invest significantly less In on-the-job training: longer job duration Is associated with steeper growth In earnings.

2. Although job mobility Is associated with significant wage gains (across jobs), there Is a substantial wage differential between the mobile and the non-mobile at older ages. At young ages, however, no wage differential was detected, thus the gains from mobility plus the lower costs of Investment are compensating the lower returns accruing to the mobile from their lower on-the-job training.

As the men age, less mobility Is undertaken so that the gains from mobility fall, and the non-mobile begin collecting returns on a large volume of training, leading to higher earnings.

3. The explanatory power of the equation was significantly Increased by accounting for the effects of job mobility; this Increase occurred when the longest job ever (regardless of when It occurred) was Introduced In the equation. This Is due to the fact that most of the human capital Investment takes place In the longest job. The Increase In explanatory power, therefore, points out that job mobility Is an Important determinant of the wage structure.