4.3 Comparative Advantage and Learning Within and Between Firms
In the analysis so far, I have considered the sample of workers remaining with their firms. As
a result, the bj θi term representing the quality of the match worker-rank is defined as being
firm-specific. It would be interesting to see whether the unmeasured θ term that is driving
the results is transferable across firms (that is, individual rather than firm-specific). Using the
information on job changes, I re-estimate the model over the sample of workers moving within
and between firms. Results are presented in Table 6.
Results are similar to those obtained using the sample of firm stayers. The similarity in
the results over the two samples suggests that the unmeasured quality of the match between
a worker and his rank would be more individual than firm-specific. Note that the analysis is
based on survey data which do not allow one to identify firms and therefore to control for firm
unobserved heterogeneity. Having matched employer-employee data with similar information
on job mobility and hierarchical job structure as in the GSOEP would allow for more precise
conclusions about individual versus firm specificity in the unmeasured quality term driving the
self-selection of workers into ranks.
5 Conclusion
In this paper, I have analyzed the relative importance of different factors explaining the dy-
namics of wages and workers mobility within firms with data from the German Socio-Economic
Panel survey over the years 1986 to 1996. Using survey data for a large sample of workers
within their firm, I can draw conclusions on the common features arising from the relationship
between workers’ career inside the firm and wage outcomes over a large sample of firms. In
addition, the longitudinal aspect of the data allows me to study wage and mobility dynamics.
A preliminary analysis of the data shows that, similarly to the U.S. findings on career and
wage outcomes inside firms, average wage growth at promotion is higher than without promo-
tion. At the same time, individual differences in wages seem to remain important within each
level of the job hierarchy as average wage growth at promotion is lower than the percentage
difference in average wages between two job levels. There is also some evidence that wage
growth predicts promotion as average previous-period wage growth is higher for those experi-
encing a promotion (or a change in job) the following period than for those who do not. On
the other hand, the data show that German workers experience very few episodes of successive
mobility within the firm over the sample period.
24