The results presented earlier for those Individuals who were least mobile (Pattern 1) Indicated they Invested heavily on the job as expected since these men currently receive returns on all training ever acquired (net of depreciation), and since they had more Incentive to Invest larger amounts in their only job.
For the individuals in Pattern 2—where the first Job was a short job different from the current (longest) job—Investment was also extensive. The estimate of kQ^ for the current job Is higher than the estimates for previous jobs, despite the fact that the current job started 15.A years after the beginning of labor force experience.
The estimates for individuals In Pattern 3—where the longest job was the first job—are highly dependent on the functional form of the earnings function. One reason for the Instability of the coefficients might be the small size of this mobility pattern (113 observations). The results do indicate little Investment in all jobs.
The results for the most mobile individuals—Pattern A—show that little Investment occurred In all jobs except the longest. Both the first and current jobs yield estimated k^’8 which are negative even though In the actual regression the current job coefficient was significantly higher than all the other coefficients. The fact that these estimates are negative might be because these are ratios net of depredation.
The projected shown In Table 5 yield two important empirical findings.
Note that since we have assumed the same rate of decline within all jobs, the difference between k^ across jobs gives the shift In the level of the investment profile from one job to the other. The estimates In Table 5 show consistently higher k^ the later the job. This finding Indicates that there Is an upward shift In the level of the Investment profile after changing jobs. A second finding given by the kQl of Patterns 2 and 4 la that the size of the Juap In Investment across Jobs Is highest when Individuals switch to the longest Job. r Summarizing, two important conclusions can be Inferred from the analysis of investment ratios across Jobs: First, the results indicate that the non-‘й mobile Invest in time units more than the mobile. This, of course, depresses current earnings of mobile men, ceteris paribus; thus providing an explanation as to why the current earnings of mobile individuals are lower than the earnings of the non-mobile. Secondly, the findings show that longer duration of Jobs Is associated with higher growth In earnings.