3. A new productive system and its workforce
- 3.1 Old jobs and new
- 3.2 The demography of labour supply
- 3.3 Education, youth employment and unemployment
- 3.4 Parallels and explanations
- 3.5 'Early retirement' unexplained
- 3.6 Skills and occupations
3.1 Old jobs and new
From the late 1960s to 2000, the more apparent trends were the cumulative increase in female labour supply; more part-time employment and employment in service industries; the decrease in male labour supply; relatively slow growth in full-time employment, and the reduction in the importance of industries preparing, treating or converting materials and producing goods. These changes are so closely interrelated that priority can hardly be assigned to any, although the changing industrial structure must have been influenced by technological developments. Similar trends occurred in other industrialised countries.
Figure 1 shows the distribution over time of the Australian working population between 'production industries', comprising construction, agriculture, forestry and fishing, manufacturing, mining, and electricity, gas and water; and 'service industries', comprising property and business services, accommodation, cafés and restaurants, cultural and recreational services, personal and other services, health and community services, retail trade, education, wholesale trade, government administration and defence, finance and insurance, transport and storage, and communication services.
The rate of change in the distribution of employment between those two broad categories was remarkably steady through a period of more than thirty years, varied only by episodic accelerations and decelerations apparently corresponding to upturns and downturns of the business cycle. According to a fairly standard interpretation of this particular change in the labour force 'over the past few decades', offered by the ABS:
…new technology, microeconomic reforms (such as tariff reductions, industrial relations reforms and changes to standards and regulations) and internationalisation of product markets have all contributed to this change. [A] major influence on industrial restructuring has been government economic policy. [I]nitiatves in the 1980s aimed at creating a more efficient and less sheltered manufacturing sector…included gradual cutting of tariffs, floating the Australian dollar, liberalisation of foreign investment and allowing foreign banks into Australia… Factors contributing to…growth of employment in the service sector…include the increase in part-time and casual work and the increase in services that replace work previously done in the home [and also] the recognition that many sectors of the service industry have the potential to earn export income (ABS 1997, 'Changing industries…').
That account emphasises the significance of policy change, and perhaps of changes in tastes. However, if policy decisions were of primary importance one would expect a less regular pace of change, with rates responding perceptibly to rounds of tariff cuts or financial deregulation. The regular trend towards growth of certain industries and the decline of others could correspond to the diffusion of a new technological style, or style of life, or both. Thus the policy changes mentioned in the ABS publication could then be examples of institutional adaptation to an emerging style rather than causes of it. The possible indications in Figure 1 of a slight levelling out in the 1990s could represent the impending completion of a structural shift.
Figure 1: Proportion of all employed persons in production and service industries, 1966-1997
Redrawn from ABS 4102.0; 1997 figures 6230.0
Firstly however, as discussed later, workers providing business services may or may not be classified to the employers' industry, depending on whether the services are produced in-house or bought in.
Secondly, an industrial classification relates more directly to what is produced than to how it is produced, and emergence of a new productive style must involve more than changes in relative demand for existing goods and services, and hence for those who produce them. By themselves, productivity gains will tend to reduce demand for labour, as they have in agriculture and some sectors of the transport industry.
The specific effects of information technology on employment are not clearly apparent. For example, they will have influenced changes in finance and large-scale retailing, but much less in the construction industry (which increased its share of jobs between 1985-86 and 1995-96), or in 'accommodation, cafés and restaurants' (where employment grew at three times the overall rate in the same period). Some of the recognised industrial categories are heterogeneous: for example, 'business services', a subdivision of 'property and business services', contains both 'computer services' and 'security and investigative services (except police)', both of which grew from minor to substantial sources of employment but for quite different reasons. Any effects on demand for clerical labour could involve many categories.
The rise of employment in service industries was associated with another of the long-term trends suggestive of a historic transitionæthat is, growth of female employment. In 1997, women occupied 43 per cent of all jobs. Figure 2 shows industry subdivisions with more than that proportion, accounting between them for 76 per cent of all women workers. Except for the declining 'textile, clothing, footwear and leather manufacturing', they are all service industries. Perhaps women had some advantage in the expanding industries, or perhaps the availability of women workers enabled them to expand, or perhaps both are true. However, the subdivisions with the highest proportions of women, although certainly not untouched by technical innovation, and some of them much concerned with information, are characterised more by labour-intensity.
Figure 2: Industry subdivisions with more than 43 per cent female workers, 1997
Source: ABS 6230.0
Note: These subdivisions included 76 per cent of all employed females and 41 per cent of employed males
Between the early 1970s and 2000, employment grew approximately in proportion to the Australian population. Figure 3 shows the numbers of males and females in full-time and part- time jobs. Cyclical change was most apparent in male full-time employment, less in female full-time and least in part-time employment. All categories grew steadily throughout, although male part-time numbers came from a very low base. In Figure 4 the same categories are presented as proportions of all jobs in the particular year. As with the growth of employment in service industries, the striking feature of the trends over twenty-nine years is their regularity. Male full-time employment steadily lost its dominance, while part-time male and, especially, female grew; and female full-time employment continued to constitute around 24 per cent of all jobs. This meant many more jobs at the end than at the beginning of the period, and a much-increased proportion of all full-time employment, from 27 per cent in 1972 to 34 per cent in 2000. The statistical series again provides an image of a system in transition, but also possible indications that rates of change eased in the long period of growth following the recession of the early 90s.
Figure 3: Number of employed persons by full or part-time and sex, 1972-2000
Source: ABS 6101.0, 6203.0
All figures are for August
Figure 4: Distribution of employment by sex and hours, Australia, 1972-2000
Source: ABS 6101.0, 6203.0
All figures are for August
3.2 The demography of labour supply
In Figure 5 the labour supply of males and females from 1975 to 2000 is represented as aggregate weekly hours of paid employment divided by the population 15 years of age and over. That base seems as good as any. 'Working age' has conventionally been regarded as 15 to 64 years for males and 15 to 59 (until 1995) for females, the upper limit being age of eligibility for statutory pension, but the reality was always less distinct and has varied over time. On the present calculation, adjusted aggregate hours fluctuated around 20 or 21 hours per week for the whole period, with movements above and below that figure corresponding to phases in the business cycle. However, the index of market sector productivity superimposed on the figure indicates that the real product of that labour supply rose by more than 30 per cent.
On average, employed males continued to work longer hours than females. One reason was their concentration in full-time employment. Although the ABS classified anything under 35 hours a week as part-time, the average was much shorter—about 17 hours in August 1999, when part-timers constituted 26 per cent of all workers but contributed only 12 per cent of all hours worked. Secondly, on average, women's full-time hours of work were consistently shorter than men's hoursæfor example, 36.7 and 40.9 hours respectively in June 1989, and 39.5 and 44.2 in August 1999.
The net result shown in Figure 5 was that although male labour supply was sensitive to both downturns and upturns of the business cycle it trended down, from about 15 hours per head of total adult population to about 13 hours, a decline almost balanced by a steadier rise in the female contribution from about six to above seven hours. Again, though, the male contribution stabilised in the 1990s.
Figure 5: Hours worked per head of population 15 & over, by sex, 1975-2000, and real gross product per hour worked, 1978-1996
Source: ABS 6101.0, 6203.0
All figures are for August
Figure 6 and Figure 7 show employment ratios by age group for females and males respectively for 1972, 1979, 1985, 1991 and 1997, and female ratios also for 1961. Employment ratio (the proportion of a group in the population actually employed) is used here and elsewhere as a more robust statistic than employment rate (the employed as a proportion of those in the labour force) which varies according to economic and social conditions.
The female ratios are more readily understood than the male. They expanded for most age groups, particularly in the range corresponding to the bearing and raising of children. The 1961 figures indicate that marriage and childbearing was commonly associated with permanent withdrawal from paid employment. In 1966, only 30 per cent of married women under 55 years of age, with or without children, were in paid work. By 1997, however, the inflection now representing only a temporary withdrawal from employment had almost disappeared. Further detail is shown in Figure 8. Between 1981 and 1997, when the employment ratio of married mothers whose youngest or only child was 10 but under 15 years of age rose from 55 to 69 per cent, the ratio for mothers of children under five years of age rose even more strongly from 30 to 46 per cent.
Meanwhile, Australian women were becoming more available for employment because they were spending less of their lives with responsibility for children; young women because first births were occurring later and older women because completed families were smaller. Natalie Jackson, writing in 1998 about the cohorts of women born in 1933-37, 1953-57 and 1970-1980, noted that 'In the 1960s only 8 per cent of women remained childless [but] current fertility patterns indicate that about one-quarter of women currently of reproductive age will remain childless,' and that 'successive cohorts of women are reentering the workforce with children of ever-younger ages', and summarised:
For the cohort born 1933-37… entry into the labour force was largely completed by age 16. Marriage and childbearing followed in quick succession, with the average age at first birth being between 23 and 24 years, and childbearing largely completed by age 31… For the [1953-57 cohort] there was a small upward shift in average age at labour force entry and first birth and, largely due to smaller average family sizes, a slight decrease in average age at last birth…
By contrast, for those born in the 'baby bust' years of the 1970s… there is every reason to suspect that significant demographic compression is occurring. [L]abour force participation at the younger ages has fallen… The median age at first birth… has similarly shifted upwards to 29.2 years. Given that this was the average age at the last birth for the cohort born 1953-57, the last birth age… will have to be around 35 years (allowing the same childbearing duration) (Jackson 1998).
Figure 6 suggests that the transformation of female employment proceeded rapidly in the 1960s, and also that a considerable expansion occurred between 1985 and 1991 for women aged roughly 35 to 54. Changes in employment ratios for unmarried women, with or without children, are shown in Figure 9. After gradual decline, ratios for 55-59-year-old women trended up from the late 1980s. Comparison of employment ratios for 1991 and 2000, as in Figure 10, indicates little or no change for males and slight growth for women in the years most closely associated with childbearing, but continuing gains by older women. The relatively small changes for males and younger females are again consistent with stabilisation.
Figure 6: Female employment ratios by age, Australia, 1961-1997
Source: Labour Report 1964, ABS 6101.0, 6203.0, Rept of Comm. on Econ.
Figure 7: Male employment ratios by age, Australia, 1972-1997
Source: ABS 6101.0, 6203.0
All figures are for August
Figure 8: Married women: Employment ratios by hours and age of youngest child, 1981-1997
Source: ABS 6224.0
Figure 9: Employment ratios of unmarried women, with or without dependants, by age, 1980-2000
Source: ABS 6203.0
All figures are for August
Figure 10: Employment ratios by age and sex, 1991-2000
Source: ABS 6203.0
All figures are for August
The changes in male employment ratios from 1972 to 1997, shown in Figure 7, indicate that the steepest decline, much of which occurred in the first half of the period, was among older men, and especially men 60 and over. Year-by-year detail is shown in Figure 11. As in other series, the ups and downs of the business cycle are superimposed on long-term trends, which are downward here. The oldest of the three age groups was affected by a policy change made long before, in 1935, when the Australian Soldiers' Repatriation Act was amended to provide 'service pension', equivalent to civilian age pension but payable at age 60 instead of 65, to war veterans. In the 1970s, increasing numbers of veterans of the war of 1939-45 were turning 60 and claiming pension, until in 1983 a peak was reached, followed by rapid decline in new grants. Perhaps expectations had changed, because although the employment of men aged 60 to 64 recovered quite strongly in the 1990s—in an unusual reversal of a trend—it was coming back from a very low base. The employment ratio for men aged 45 to 54 appears to have stabilised, and the rate of decline for men aged 55 to 59 to have eased.1
However, male employment fell at all ages and for groups whose rates had once been extremely high. For example, between 1977 and 1997, while the employment ratio for married mothers of dependent children rose from 40 to 59 per cent, the ratio for married fathers fell from 95 to 88 per cent. In Figure 12 and Figure 13 data from the quinquennial census are used to reconstruct, somewhat approximately, the experience of men born in successive five- year periods from 1897 to 1976. For ease of reading each of the figures shows alternate age cohorts. The strong shift at older ages is again apparent, clearly beginning with the 1912-16 cohort when it reached its early sixties and affecting the 1917-21 cohort in its late or even its early fifties, but this presentation reveals a new feature: not only does each successive cohort's employment follow a lower trajectory but the peak tends to have been reached at younger and younger ages. The 1937-41 cohort seems to have been at its best in 1971, when aged 25 to 29. The slight inflections apparent between the last and third-last data points for the 1952-56 and 1957-61 cohorts are probably because of high cyclical unemployment at the time of the 1991 census. Figure 10 suggests a stabilisation in the 1990s.
Tendencies for increasing female labour supply to compensate for declining male labour supply did not necessarily extend to the family and household. Figure 14 shows proportions of couple families with dependent children from 1979 to 2000 in which husband, wife, both or neither was employed. The main trends were increase in the proportions of two-earner families (which became the majority) and families with no earnings. The wives of non-employed husbands were not more but less likely than other women to be working. The proportion of families with no earnings increased at each major downturn and recovered only slightly but appears, like the ratios for other family types, to have stabilised in the later1990s.
Figure 11: Male employment ratios by age group, 1972-2000
Source: ABS 6101.0, 6203.0
All figures are for August
Figure 12: Male employment ratios by birth cohort, 1961-1996
Source: Published data from quinquennial censuses 1961-1991, 1996 data supplied by ABS Alternate 5-year cohorts omitted
Figure 13: Male employment ratios by birth cohort, 1961-1996 (alternate cohorts)
Source: Published data from quinquennial censuses 1961-1991, 1996 data supplied by ABS Alternate 5-year cohorts omitted
Figure 14: Couple families with dependent children: Employment of husband and wife, 1979-2000
Source: ABS 6224.0
Single parents constituted another example of a family type where lack of adequate earned income was associated with lower rather than higher labour supply. As shown in Figure 15, between 1977 and 2000 employment ratios for single parents were lower than for married parents, and the differences increased over time. Figure 16 distinguishes rates of full-time and part-time employment for single and married mothers from 1983 to 2000. Rates of increase of part-time employment were similar, except that the married rate may have fallen slightly in the late 1990s. In the earlier part of the period, single mothers were more likely to work full time than married mothers but their full-time employment actually declined from about 1991— another unusual reversal of a trend—and became less common than part-time, while an increasing proportion of married mothers worked full time until, again, the rate stabilised.
The employment ratios of both males and females aged under 25 years, and especially under 20, declined because of reduced opportunities for full-time employment and rising rates of participation in secondary and tertiary education but, because of growing opportunities for part-time employment and increasing overlap of the population of full-time students and the adolescent and young adult labour force, less than they might otherwise have done. Public discussion of youth unemployment was notably confused by failure to take account of the nature and extent of the changes.
Figure 15: Parents of dependent children: Employment ratios by sex and marital status, 1977-2000
Source: ABS 6224.0
Note gaps in series
Figure 16: Employment rates of married and single mothers, 1983-2000
Source: ABS 6224.0, 6203.0
Figures July to 1985, then June
3.3 Education, youth employment and unemployment
At the 1971 census, 1.5 per cent of the male workforce was unemployed. Participation rates for males aged 15 to 19 and 20 to 24 were 56 and 89 per cent respectively and the unemployment rates 3.8 and 2.1 per cent. In August 1998, the male unemployment rate was 8.3 per cent, the participation rates of males aged 15 to 19 and 20 to 24 were slightly lower than 27 years before, at 54 and 86 per cent respectively, and the unemployment rates were many times higher, at 20.6 and 12.8 per cent, but so much else had changed that these and the 1971 figures were simply not comparable.
In the 1980s and 1990s, adolescent and young adult rates of participation in secondary and tertiary education rose to unprecedented levels, and females made greater gains than males, with probable enhancement of their future earning capacity. Figure 17 shows rates of employment and full-time study of males and females aged 15 to 19 from 1987 to 2000. The effects of sharp downturn in the business cycle in 1990 with a low point in 1992 followed by recovery are apparent in each of the series, with indications of a reciprocal relationship between employment and retention rates. Male and female levels of employment were similar and moved together, but females achieved somewhat higher rates of participation in education. Figure 18 shows that similar growth occurred and similar differences emerged among 20 to 24 year-olds, the rising trend being stronger for females than for males, with male rates of participation in full-time tertiary study moving from 11 per cent in 1987 to 20 per cent in 2000 and female rates more than doubling, from 9 per cent in 1987 to 23 per cent in 2000.
The educational system is only one of several means of replenishing and expanding the stock of human capital, although it has tended to displace or absorb some of the others, and in particular various forms of in-service training. Traditionally, apprenticeships in skilled trades (mostly available to males) were an alternative to continuing in formal education. Freeland, whose concern was with teenage unemployment, referred to the apprenticeship system as 'a protected labour market reserve for young males making the transition from school to employment', and called for 'an equivalent training culture across all industries' (Freeland 1997, pp 29-31).
Apprentices will normally be classified as employed, and in 1997 were 90 per cent male. Figure 19 is intended to give some idea of how the broader trend for adolescent males looks when apprentices are added to students. (As noted on the diagram, figures for apprenticeships are approximations, required by lack of detail in statistics for the earlier years, but they serve as an indicator of volume of entry.) The conclusions supported are that a fairly constant proportion of teenage males went into apprenticeships, although numbers were affected by the business cycle, and that a great and growing majority was in either formal education or apprenticeships, but that females had made real educational gains.
The last years shown in Figure 19 understate participation in formal training because increasing numbers of young people were entering traineeships. In 1996, apprenticeships and traineeships were integrated as 'new apprenticeships' intended, among other things, to provide better training opportunities for female adolescents and young adults.
In August 1997, 18 per cent of all part-time workers and 28 per cent of male part-time workers were full-time students under 25 years of age. In 2000, 53 per cent of employed persons under 20 years of age were full-time students, and 30 per cent of male but 41 per cent of female students were employed. Among 20-24-year-olds, as in the younger group, female students were more likely to be employed than males, with rates of 54 and 45 per cent respectively. Figure 20 shows a continuing upward trend in young-adult student employment.
Figure 17: Persons aged 15-19: rates of employment and full-time education, 1987-2000
Source: ABS 6203.0
All figures are for August
Figure 18: Persons aged 20-24 in full-time tertiary education, 1987-2000
Source: ABS 6203.0
All figures are for August
Figure 19: Males 15-19 in full-time education, and all apprenticeships current, 1987-1997
Source: ABS 6203.0, 6227.0
Figures for apprenticeships include females and persons 20 and over
Figure 20: Full-time students aged 20-24: percentage also employed, by sex, 1987-2000
Source: ABS 6203.0
All figures are for August
In a review of trends from 1966 to 1995, Wooden describes the 'enormous change' that had occurred in the youth labour market and notes the rapid increase in part-time employment but 'a decline in the number of full-time jobs…and a rise in youth unemployment' (Wooden 1996). That is true, but by the 1990s the term 'unemployment' had not only changed its meaning but had come to refer to a variety of situations hardly comparable with each other.
In August 1998, unemployment, by the standard definition, was 18.8 per cent for persons aged 15-19 and 11.9 per cent for persons aged 20-24 but, of course, the percentages are calculated on the base of persons working or actively seeking work, and 46 per cent of the teenagers were doing neither, mainly because they were in full-time education. But nevertheless, 46 per cent of the unemployed were full-time students, of whom 84 per cent were looking for part-time work. It can't be said that they didn't need that employment, but they hardly depended on it for a livelihood, and a definition closer to what we think we mean when we talk about adult unemployment might be 'not a full-time student, not working and seeking full-time work'. That would give a figure not of 19 per cent but 5.2 per cent of teenagers. In 2000, the corresponding figures were 16 per cent and 3.4 per cent.
Rates calculated on that limited basis are shown in Figure 21. In the 14-year period, male teenage unemployment rose and fell between 4 and 9 per cent, and for females, between 3 and 7 per cent. The calculated trend lines indicate falling rates from 1992, which is interesting because this dwindling minority of young people who had discontinued education could have suffered disadvantages in a presumably more demanding labour market. In the 20 to 24 year age-group male rates were much higher, moving between 6 and 15 per cent, although from 1993 the trend was down. Female rates moved only between 4 and 8 per cent, but Figure 21 shows another complicating factor—the high but variable proportion of non-student young women classified as not in the labour force. The calculated polynomial trend indicated falling rates to 1994 and then rising. No satisfactory explanation is apparent.
However, unemployment rates for young people in the full-time labour market were very high. In August 1999, of the 21 per cent of teenagers and 59 per cent of young adults either in full-time work or unemployed and seeking it, 20.9 and 11.6 per cent respectively were unemployed. Wooden (1996) concluded from his review of trends that 'the future labour market experience of today's youth cohort' would probably differ from that of its predecessors, but in ways and to a degree so far unpredictable. They might benefit from their extended education, 'or will the higher levels of exposure to casual and intermittent employment inhibit skills development and long-term career prospects?' Then again, 'there are good reasons to be concerned about those young people who exit the education system at a relatively young age and find themselves unable to secure full-time employment.' He did not go on to consider possible differences between males and females, but it seems that in some important ways females may have the better prospects, and that most of those who suffered from any polarisation would be male.
Figure 21: Non-student, not employed full-time jobseekers by sex and age, and females aged 20-24 not student and not in labour force, 1987-2000
Source: ABS 6203.0
All figures are for August
3.4 Parallels and explanations
In general terms, and sometimes in detail, Australia underwent the same changes as countries with comparable economic systems, including its principal trading partners. Similar trends in the industrialised and industrialising countries are evidence of the depth of the process, and differences in timing, form and degree demonstrate the influence of local contexts, both social and institutional. The Jobs report published by the Organisation for Economic Cooperation and Development (OECD) in 1994 noted that:
A number of labour market trends were remarkably common among all or most OECD countries… A common feature of all OECD countries has been a slowdown in growth in both output and productivity over the 1980s and 1970s as compared to the 1960s… In the early 1990s more people were unemployed in the OECD area than at any time before… From 1973 to 1982 [the] number [had] tripled. Strong economic expansion over the latter half of the 1980s did bring some reprieve but failed to drive down the jobless total to the levels of the 1970s, much less to those of the 1950s and 1960s… (OECD 1994a).
Unemployment in several member countries was to rise higher still in the later 1990s. The 1994 study also reported that:
…growth of employment in the service sector over the 1980s, relative to the agricultural and industrial sectors, was…a feature of earlier decades… However…prior to 1970, the decline in the relative importance of agriculture was compensated by increasing employment shares in both the manufacturing and service sectors. Thereafter, the relative importance of the industrial sector as a source of jobs has also declined… Not all of this shift is entirely due to changes in consumption patterns, trade and technology; it also partly reflects a growth in the contracting out of service activities… In virtually all countries, employment expanded the fastest in financial and business services and in community and personal services… Another feature of the 1980s was the growth of part-time work in many OECD countries. In fact, in several countries, notably Belgium, Ireland, France and New Zealand, the number of full-time jobs declined over the 1980s and it was only through a rise in part-time jobs that overall employment gains were achieved. Part-time work also accounted for almost all of the overall rise in employment in Germany and the United Kingdom… The Netherlands experienced a particularly rapid rise…
One common trend across all OECD countries over the last 15 to 20 years has been the decline in participation rates for men and the rise for women (with the exception of Turkey). Outside the European Community, the increasing numbers of women entering the labour force have more than offset the decline in men's activity and, consequently, overall participation rates have been growing over time, with the biggest rise occurring in North America… Increases in non-employment rates for older men have been particularly substantial…
Non-employment rates for young people have increased substantially [except in] Germany, Japan, the Netherlands and the United Kingdom. However, this partly reflects higher retention rates of youths in education as well as increases in unemployment rates (OECD 1994a, pp 1, 2, 5, 8-9, 28-33).
Another study published by the OECD in the same year dealt specifically with trends in female employment. Expansion of the service sector had favoured the employment of women, service sector firms had been 'the keenest proponents of…external numerical flexibility [and] there is one form of flexibility that is considered particularly "female", namely part-time employment [which] has the undeniable advantages of creating new employment opportunities for many women.' In some countries a further incentive for employers to offer part-time work was that their contributory social insurance schemes excluded persons working short hours. 'Short' part-time work was known to have increased rapidly in several countries. 'The proportion of men working part time is still uniformly low [and] men working part time, especially on "short" hours, are typically students in Denmark, the Netherlands and the United Kingdom.' Denmark was unusual in that the great majority of women working 'short' hours were also students (OECD 1994b, 47-49, 78-9, 94).2
3.5 'Early retirement' unexplained
Declining male participation and labour supply aroused concern to the extent that it was reflected in unemployment and 'early retirement', but little curiosity, and was not studied closely. Two suggested explanations, not mutually exclusive, were that structural change reduced opportunities in traditionally male occupations and industries—or more generally resulted in mismatch between opportunities and male skills and preferences—and that a higher proportion of men were able to retire voluntarily. As expressed by the ABS, 'The production industries were, and still are, mainly the preserve of men working full time, and in the 10 years from 1985-86 to 1995-96 jobs in them increased by 0.7 per cent only, but service industry jobs by 31.1 per cent (ABS 1997).
The OECD's Jobs study reported, 'In all countries for which data have been collected, participation rates have fallen the most for adult men with few educational qualifications' (OECD, 1994a p. 31). A study of male participation rates from 1971 to 1993 conducted by the United States Bureau of Labor Statistics found a:
…continuous drop in labor force participation rates among men aged 25 to 54… For men aged 25 to 29 and 30 to 34; the period of the greatest drop in… participation was 1978-83—the years following the most rapid increase in the labor force. This may reflect the difficulties of the baby-boom generation getting jobs…
An important element in projecting future labor force growth is to have a better understanding of why men's labor force rates are declining. For young men, one reason …has been their increasing college enrolment. However, for men 25 to 49 that cannot be a significant factor . [Although] the number of men 25 to 49 grew by 2 per cent annually between 1983 and 1993 …the number of men…reporting themselves [in the Current Population Survey] unable to work grew by 9 per cent a year or four times as fast…
Accumulating evidence indicates that the decline in labor force activity…is greatest among men with fewer years of educational attainment and among men with low rates of pay (Fullerton 1995).3
In Australia too, many of the males whose predecessors would have been in the labour force, especially the older, classified themselves or were classified as 'unable to work' because of sickness or disability, although they were living longer and were presumably healthier.
The second line of explanation is represented by Rima, also writing of the United States, who concentrated on older men. In 1959, the labour-force participation rates of both black and white males had been:
…just under 84 per cent. However, since the early 1960s that has no longer been the case. The participation rates…declined by 1990 to 77 per cent for white males, and to 70 per cent for black males… The causal factors…are varied and complex, but the secular rise in real wages and earnings that has accompanied economic growth is the most obvious among them. [M]en as a group have chosen earlier retirement, partly in response to their having accumulated more wealth than was possible for earlier generations of males… The growth of private pensions, combined with a system of social security benefits that severely limits earnings from part-time employment…also has [increased] the attractiveness of labor-market withdrawal (Rima, 1996 p. 149).
A variant of the same theory was proposed in an Australian study of trends to 1982, which concluded that:
…the decline in participation rates over the 1970s is the result of the realisation of plans formed prior to 1973 [and] the deterioration of the labour market in the late 1970s …should be heavily discounted. [These] changing retirement plans are largely reflections of…the tendency of individuals with war service to retire at age 60 [and] change in individual preferences which derive from the greater lifetime income [and] wealth of today's 60-64-year-old cohort…(Miller 1983).
That explanation may have had some validity at the time but by the late 1990s seemed inadequate, the survivors of the large ex-service cohort having passed into old age and any plans formed before 1973 having been exposed to quarter of a century of new realities. Although home ownership and children's independence may help older people adjust to reduced income, the notion that accumulated 'wealth' enabled many to choose early retirement seems inconsistent with the fact that most were on unemployment and disability-related social security payments. The very term 'early retirement' is probably misleading when participation and employment rates had fallen for younger as well as older men, though not as far.
Borland used ABS material to examine changes in the labour-market participation of males of all ages from 15 to 64 between 1964 and 1994. Having noted that 'the issue of trend movements appears to have been relatively neglected compared to…cyclical movements in labour force participation', he found that 'The increase in the proportion of men out of the labour force was concentrated among older men, men with the lowest levels of educational attainment, with potential earnings in the middle of the distribution of earnings, and who were in married family groups.'
Like Miller, Borland compared surveys of retirement intentions with the subsequent experience of the populations surveyed, but found that 'retirement intentions do not match the decrease in…participation that occurred during the 1980s [and] it seems reasonable to conclude that contemporaneous changes in the labour market variables, as well as demand constraints, are likely to have affected…participation' (Borland 1995). However, in 1996 EPAC suggested that 'Rising personal incomes and wealth (often from having a wife as a second income-earner) have enabled older workers to retire earlier,' and that a further factor was the spreading coverage of occupational superannuation (EPAC 1996b, p. 45).
An American study estimated that changes to pension plans and social security entitlements in the 1970s and 1980s were responsible for no more than one-quarter of the trend to early retirement, and suggested that 'other factors which may have played a major role in increasing retirement include rising real wages, changes in disability insurance, changes in tastes for retirement and changes in the industrial distribution of jobs' (Anderson, Gustman & Steinmeier 1997).
Both the 'involuntary exclusion' and the 'escape to affordable leisure' explanations are compatible with the idea that secular decline in male employment was related to deep structural change and each may have some validity, but neither seems adequate. For example, many of the production workers shed by uncompetitive industries would have had few transferable skills and, although not wealthy, would not have been thrown into abject poverty by inability to find new jobs but could continue to live much as before, if at a lower level. Also, the customary and formal privileges males enjoyed in recruitment, retention, promotion and pay began to erode in the 1960s, and if men and women were competing in the labour market and within enterprises the competition was less unequal. The greater inclusiveness of the labour force and technological change could conceivably have raised the threshold of employability in terms of basic competences if not of specific skills.
Theories would need to take account of the duration of the process of change and the involvement of males of all ages. It might seem plausible that the combination of structural change and periodic downturns of the business cycle would remove workers who had become marginal and prevent their return to employment. However, the sharp rises in unemployment that occurred in 1974-75, 1981-83 and 1990-92 cannot be related clearly and consistently to the movements in employment ratios discussed above. On the contrary, we have seen possible evidence of recovery of cohort ratios following recession, and many of the new jobs were not conspicuously demanding either of skills or general competence. And there had been a whole generation for the actual and potential labour force to adapt to change.
A 1999 Treasury paper on trends in labour force participation, observing that 'the male participation rate has been declining over the last thirty years, while the female…rate has been rising,' offered the familiar explanations of employment growth in 'part-time service industries, which are dominated by females, and slower growth in industries traditionally employing full- time male workers, such as utilities, mining and manufacturing,' as well as increased female labour supply (Department of the Treasury 1999). The description is accurate enough except that, to some unknown extent, changes in industrial organisation and contracts of employment had reclassified jobs from 'production' to 'service', and the explanation implies a peculiar male inability to adapt, persisting for a whole generation. If type of work is the crucial factor, is it supposed that males have an ineradicable affinity with machines? Or if hours of work are crucial, that the labour market can't accommodate that preference as well as the preferences of women with domestic responsibilities?
3.6 Skills and occupations
A study of trends in the United States between 1959 and 1995 (Carnevale & Rose 1998) took 'a new approach because the traditional practice of tracking employment trends by industry…misses activities that cut across industries or are far removed from direct production or service activity.' Occupational data were classified into 'a new set of categories based on workplace functions', namely, extractive production, industrial production, low-skilled services, high-skilled services and 'administration and coordination (the office): all workers involved in management, administration, business and financial services'. A further classification distinguished 'élite jobs' performed by 'managers and professionals', 'good jobs' and 'less- skilled jobs'. The main findings were that:
Office work…has become the dominant feature of our economy, employing 41 per cent of all workers, paying the highest salaries, growing the fastest, employing over one-half of all college graduates and capturing 50 per cent of all earnings. [B]usiness professionals are the largest component, accounting for 44 per cent of office jobs. As the Office Economy has risen, the industrial economy has fallen…The…share of low-wage service jobs in the economy has remained constant at approximately one-fifth of the total workforce [and] high- skilled services (e.g., education, health care, police and firefighters [and clerks, technicians, tradesmen and supervisors]) have become more important…From 1959 to 1969 their employment share jumped from 11 to 16 per cent and then stabilised [although] their pay…has continued to increase, and now total earnings in these activities almost equal those of the entire industrial sector.
Decline in less-skilled jobs had been offset by increase in 'élite' jobs. Most office jobs were 'élite' or 'good', and for females over 70 per cent of élite or good jobs were in offices; where as many college-educated women were now employed as in health and education.
For females, the story is one of earnings gains for all. But for males, except for college- educated workers, earnings were down from 1979. [E]conomists and politicians…worry about [the] social consequences [and many have thought that] it must be technological change that particularly benefits those with more education. At a time when computer use has been skyrocketing, there is a certain appeal to this argument…
Yet, we find that the pay of professionals in technical/scientific fields has stagnated or declined, while the pay of nontechnical, nonscientific…professionals and managers has increased. These office careers…are growing fields that are populated with workers who have bachelor's degrees. They have some computer skills but they hardly are high- technology information workers (Carnevale & Rose 1998, p. 18).
No comparable study was conducted of the Australian labour market, but there were some parallels. Cully used data from the ABS Labour Force Survey, classified according to the 1996 revision of the Australian standard classification of occupations, 'explicitly hierarchical in ranking jobs by skill level', to examine changes in occupational composition between 1993 and 1999. The classification provided 'nine major groups,…which fall into five distinct skill levels.'
It appeared that the net employment growth of 17.2 per cent had been concentrated heavily in the highest and lowest skill levels, and in particular in jobs occupied by professionals, with degree or equivalent qualification, and by elementary clerical, sales and service workers— where employment grew by 31.5 and 35.5 per cent respectively. Employment of advanced clerical and service workers had declined slightly. To Cully this suggested 'an increasing polarisation of the Australian labour market between jobs that are high skilled (high paid) and jobs that are unskilled (low paid)' although, within the lowest skill level, growth in jobs for 'labourers and related workers' was below the overall figure (Cully 1999).
Government administration consists essentially of processing information, and during the 1980s and 1990s new technology changed both organisational structure and content of jobs in the Australian Public Service. The great bulk of staff had belonged either to the 'Third Division' or 'Fourth Division'. The minimum qualification for third division entry was a completed secondary education or equivalent, and most fourth division staff performed ancillary functions, which included filing and typing. They were represented by a separate union, the Australian Public Service Association (APSA).
Since 1967, the Department of Social Security had used electronic computers, depended on them increasingly and, having made extensive use of contract programmers, reached a certain maturity in computer applications in the early 1980s. Some fourth division staff felt their jobs were threatened and reacted defensively. In September 1983—the year that 'repetition strain injury' achieved epidemic status in Australia (Hall & Morrow 1988)—departmental representatives signed an agreement with APSA that clerical staff would not be permitted to operate keyboard equipment, which was intended and understood to mean principally word processing. In 1986, the union twice protested vigorously against use of word processing logons by third division staff. It was 'a crucial issue for our keyboard members whose jobs are threatened'. The agreement lapsed with merging of the third and fourth divisions in 1987 and disappearance of the jobs that had been defended. The Department of Social Security's annual report for 1991-92 said that 'Almost all staff make extensive use of the computing systems in their day-to-day work.' In the process almost all learnt to type, well or badly, and as of 1999 the great majority of staff in the former regional office network, now Centrelink, were involved in both processing and client contact. Few were required to have degree qualifications, although an increasing proportion had. Many vacancies were advertised with the information that certain qualifications or experience were preferred.