1           Singapore has experienced a healthy rate of job growth over the years, with unemployment maintained at a low 2% in recent years. However, of greater significance to a resource-constrained economy like Singapore is the quality of such employment growth. As the economy undergoes economic restructuring, there is considerable displacement and replacement of jobs both within and between industries as companies progress up the value-added chain or relocate elsewhere.

2           This paper examines the quality of employment gains between 1983 and 1996 in terms of job earnings for an employment matrix of 6 industrial and 6 occupational categories. The resultant 36 occupation-industry cells are arrayed in order of their gross median wages averaged over 1991-95. They are then grouped into three earnings groups - highest, middle and lowest - with each group accounting for about one-third of total employment in 1990. The analysis is largely based on data from the Labour Force Survey which covers households only and excludes workers living on sites and commuting daily to and from Malaysia. It also excludes the agriculture & fishing, quarrying, and the electricity, gas & water industries, as well as industries and occupations that are not classified. Jobs in the 36 occupation-industry cells accounted for about 90% of total employment in 1990¹.

3           From an industry viewpoint, the financial and business services industry accounted for the largest share (28%) of total jobs created between 1983-96, a vastly disproportionate increase given its share of employment (8.6%) in 1983. In contrast, the manufacturing industry - the largest employer - only accounted for a modest 12% of employment gains, and saw its share of employment declining from 30% in 1983 to 24% in 1996. Underpinned by strong employment growth in the financial and business services, the services industry as a whole - which also comprises commerce, transport and communications, and community and social services - experienced an increase in its share of employment from 62% to 69% over the period. These trends towards a more service-oriented economy are consistent with developments in the other Newly Industrialized Economies (NIEs) and the more advanced economies.

4           From an occupational perspective, the managerial and professional categories accounted for a disproportionate share of employment growth. These occupational categories accounted for three-quarters of total employment gains between 1983 to 1996, compared to their combined employment share of 22% in 1983. In contrast, growth in the production and related worker occupational categories was negligible.

5           Analysed by earnings groups, three distinct phases of employment growth can be identified over the 1983-96 period: the consolidation phase (1983-86), the extensive growth phase (1986-91) and the intensive growth phase (1991-96). The consolidation phase was characterised by a decline in overall employment as the economy went into a recession in 1985, led by substantial job losses in the two lower earnings groups. During the extensive growth phase, robust employment gains were experienced across all three earnings groups as the economy recovered and gathered momentum. In the intensive growth phase, there was a shift towards higher quality jobs, as the highest-earnings group saw large employment gains while the lowest-earnings group experienced substantial job losses.

6           Indeed, over the entire period, the highest-earnings group experienced the fastest growth rate of 6% p.a., and accounted for three-quarters of Singapore's overall employment gains. The managerial and professional occupational categories mentioned above comprised the bulk of the highest-earnings group, and accounted for virtually all the employment gains. Of these, two-thirds went to managers and professionals in the services industry, led by financial and business services. Managerial and professional employment in manufacturing also grew strongly, accounting for about a quarter of the job increase in the highest-earnings group over the period, even though manufacturing employment as a whole grew by much less. Although there was a wide dispersion in earnings within the highest-earnings group, employment growth was evenly distributed within the group, not skewed towards either the upper or lower end of the earnings scale.

7           With employment growth largely concentrated in the highest-earnings group, employment in the middle-earnings group experienced modest growth of only 1.7% p.a., mostly in the clerical and service & sales jobs in the services industry. The lowest earnings group saw hardly any growth, and has experienced a trend decline in employment since 1991. This has been the result of substantial job losses in the manufacturing industry, comprising primarily production and related workers. That the manufacturing industry has been adding managerial and professional jobs while at the same time shedding production and related jobs is indicative of its restructuring towards more capital-intensive and higher-skilled activities.

8           In line with the rising quality of employment, the aggregate skill level of the labour force has improved. The proportion of skilled labour - defined as workers with at least post-secondary education - doubled from 15% in 1983 to 31% in 1996. In contrast, the ratio of unskilled labour - those with at most lower-secondary education - declined from 54% to 39% during this period. The improvement in workers' educational level was broad-based across all industries, with the manufacturing industry experiencing the largest improvement.

9           A good part of the improvement in the economy's aggregate skill level was due to the so-called within or occupational mix effect (90%) rather than the between or industry shift effect (10%). Occupational mix effect arises from the substitution of more educationally qualified workers for less qualified ones within individual industries due to: (a) developments in the supply of better educated workers and the concomitant decline in their relative cost; and/or (b) the introduction of skills-biased technological progress. Industry shift effect stems from changes in the patterns of demand or the structure of the economy, which result in the higher-skilled industries experiencing greater-than-average employment growth compared to the lower-skilled industries.

10           The industry shift effect was concentrated almost entirely in the financial and business services industry, an industry with rising employment share and higher skill requirement. In contrast, the contribution from the manufacturing industry was negative, as the industry experienced a trend decline in its employment share over the period. On the other hand, with increasing proportions of skilled workers over the last 13 years, all industries contributed positively to the occupational mix effect. In particular, the manufacturing industry accounted for about one-third of the overall occupational mix effect.

11           With both its positive contribution to the occupational mix effect and negative contribution to the industry shift effect, the manufacturing industry makes for an interesting case study. Between 1983 and 1996, both value-added and capital stock in the manufacturing industry grew at roughly similar rates which were well in excess of its employment growth. The substantial increase in capital-labour ratio, or capital deepening, in the manufacturing industry reflected its shift to more capital-intensive, labour-saving production techniques and activities. These activities were also increasingly skills-biased, requiring greater use of skilled workers than unskilled workers. Indeed, since 1991, there has been an absolute decline in the number of lower skilled manufacturing workers.

   The analysis will be biased towards findings of better job quality only to the extent that the excluded workers are lower-skilled ones and that their number has increased during the period. However, estimates which conservatively assumed that all the excluded workers were unskilled ones showed the bias to be insignificant, and would not alter the findings of this paper. This paper can be downloaded in PDF format