4  Employment growth and the role of qualifications

One of the main purposes of this paper is to examine the way in which the upskilling documented in the previous section has interacted with labour market changes. The summaries of skill intensity patterns by industry and occupation have already provided some indication of changes. In this section, we examine in more detail the changes that have occurred between groups of workers defined by industry and occupation. We begin by describing how employment has been distributed across industry-occupation cells, and changes that have occurred in the distribution during the period. We then incorporate qualifications, and evaluate the relative importance of qualification, industry and occupation specific effects in understanding employment change over the period. Loosely speaking, we consider qualification effects to reflect labour supply side factors, occupation effects to reflect production side factors, and industry effects to reflect product market factors.

Our focus of analysis is on the allocation of qualifications in the labour market across “jobs”, which we define according to the interaction of the 28 industry and 24 occupation groups described above: thus generating 672 possible cells.

4.1  The changing distribution of employment

Figures 2a to 2d describe the distribution of employment across the industry and occupation cells for each of the four census years.[16] The vertical axis on each figure lists the 24 occupation groups, ordered according to each occupation’s 1986 relative skill intensity, and shows the fraction of total employment in that occupation group (and the year-specific skill intensity in parentheses). Likewise, the horizontal axis lists the 28 industry groups, ordered by each industry’s 1986 relative skill intensity, and shows the fraction of total employment in that industry (with year-specific skill intensity in parentheses). Each cell in a figure then corresponds to an industry-occupation specific job, and is shaded according to the fraction of total employment in that job: the darker cells represent the jobs with relatively large employment.

It is both evident and unsurprising that employment is not evenly distributed across the cells. This is partly due to industries and occupations having different sizes, and partly due to the distribution of occupational employment within an industry varying across industries. For example, in 1986 the largest single job cell, agricultural workers employed in the primary production industry group, accounted for 6.0% of total employment in 1986, and this share had fallen to 3.6% in 2001. The second largest cell, teaching professionals in education, had 4.1% of employment in 1986, and this increased to 4.3% in 2001. Furthermore, 29 of the 672 job cells account for half of employment in 1986, and 33 job cells account for half of employment in 2001.

Although purely anecdotal, it is illustrative to compare the changes over the period in the employment, together with the changes in the qualification distributions, of the two largest job cells just mentioned. First, agricultural workers have predominantly low qualification levels: in 1986 the largest fractions of workers had no qualifications (42%), other post-school qualifications (23%), or school certificate (16%). While employment in this job cell fell by nearly one-third between 1986 and 2001 (from 77,856 to 52,950), the skill levels increased markedly: the fraction of workers with no qualifications fell by one-third to 28%, and there was a general increase in the fractions of other qualifications over this period. On the other hand, teaching professionals are generally highly qualified: the largest three qualification groups were other post-school (57%), bachelor degrees (20%) and higher degrees (17%). Total employment of these workers increased by 2% (from 52,785 to 63,369), and this also coincided with an increase in qualifications: the employment shares with bachelor and higher degrees increased to 33% and 20% respectively, and the other post-school share dropped to 38%. Comparing the changes of these two job cells suggests that both a shift to skilled employment across job cells and also an increase in skills within job cells occurred.

The first dimension along which we explore the growth in employment over the period is by the initial (1986) size of the job cell. Figure 3 graphs the change in log employment between 1986 and 2001 against the log of 1986 employment for each industry-occupation cell, together with the least-squares fitted regression line. Two features of the relationship are apparent. First, the relationship is negative (the slope of the fitted regression line is –0.26, standard error = 0.02), implying some convergence in size and/or reversion to the mean occurred.

Second, there is clearly a great deal of variation, both in employment size and growth rates. For example, 409 job cells, employing 492,000 workers (29%) in 1986 either grew or declined by over 50 log points in the following 15 years.[17] Furthermore, in a regression of the change in log employment between 1986 and 2001, only 22% of the variation is explained by 1986 log employment. It is this variation in growth rates that we will analyse to identify a relationship between employment growth and qualifications.