6 Conclusion
We have presented a variety of analyses in this paper on the changing nature of qualifications, employment and incomes of the working age population (aged 20-59) over the period from 1986 to 2001, and how these changes may be related. First, we show a general upskilling occurred between 1986 and 2001 that is evident across a wide range of job-groups and labour force states. Abstracting from this aggregate pattern, we find that the upskilling arose due to a rise in the proportion of the population with university and school-level qualifications, and a decline in the proportions with other post-school qualifications and no qualifications.
Second, using regression-based analysis of the relationship between employment growth indicates that qualification-related factors account for around one-third of the variation in employment growth, regardless of whether industry and occupation effects are controlled for. Employment growth was strongest in job-groups that had high initial levels of skilled workers but appears to be only weakly related to subsequent upskilling. Furthermore, the relationship between employment growth and initial skill-intensity does not reflect primarily industry or occupation-related factors. The absence of a relationship between growth in skill-intensity and growth in employment weakens the case that an upskilling in qualifications causes employment growth.
Third, the value of human capital, as measured by average annual incomes, increased about 20% between 1986 and 2001. Around 75% of this increase can be attributed to general increases in incomes across qualifications at constant qualification shares, while between 15-20% appears to be due to the effects of upskilling at constant incomes, and the remaining 8% to the interaction of upskilling and increasing incomes. At the individual qualification level, the results are consistent with the notion that the upskilling that occurred appears to have been in response to increase in demand for skills.
The patterns differ somewhat for particular subperiods, as well as across different populations defined according to labour market status. These differences appear to reflect a combination of business cycle effects, which affect both the aggregate level changes as well as cause differential selection into different work force states over time, and also apparently secular changes in product and factor markets trends, such as skill acquisition.
