Labour force ageing: Its impact on employment level and structure. The cases from Japan and Australia

Transcription

1 Labour force ageing: Its impact on employment level and structure. The cases from Japan and Australia Ewa Orzechowska-Fischer The Australian National University Abstract Introduction: In recent times, the labour market consequences of population ageing have become a major concern in the population ageing debate. Often addressed in the debate is the possibility of labour force shortages and its subsequent economic implications. The study below hopes to be a contribution to this debate. The focus of the study is on examining the relationship between changes in the age structure of the population aged years and over and employment on the aggregate and industrial level. Comparing two countries (Japan-Australia) in different stages of ageing allows on the simultaneous analysis of different levels of ageing Data and method: The relationship between changing age composition and employment is addressed by the means of one factor decomposition analysis of employment in Japan and Australia for census years from 96 to. The decompositions considered age and employment rates by age as components of change in the overall employment rate across an intercensal period. The work was carried out for both males and females for all industries combined and for a classification of industries into types that are comparable across the two countries. Conclusions: Given age-specific employment patterns of particular industries, with the advances in the ageing of the labour force: - on the aggregate level, the significance of the age composition effect grows, and tends to take increasingly negative values (the values for the age composition effect are much stronger for Japan, than for Australia). Given the current age-specific employment patterns, an aged population structure becomes less favorable to the total employment - on the industrial level, the significance and direction of the age composition effect varies for particular industries. Although some industries benefit from the changes in the age composition, (i.e. Agriculture), for most (i.e. Wholesale-retail trade, et. al) the age compositional changes tend to be unfavorable. In the context of results of the above analysis, the ageing of the labour force, under present industrial structure and in the absence of significant productivity increases and policy measures, points towards labour force shortages, and economic slow down. Paper prepared for the th Biennial Conference of the Australian Population Association, -7 September 4, Canberra.

2 LABOUR FORCE AGEING: ITS IMPACT ON EMPLOYMENT LEVEL AND STRUCTURE. THE CASES FROM JAPAN AND AUSTRALIA. A paper presented at the Australian Population Association s, th Biennial Conference, Canberra, -7 September 4 by Ewa Orzechowska-Fischer Phd Candidate Demography and Sociology Programe Australia National University T: F: Ewa.Orzechowska@anu.edu.au Supervisory panel: Chair: Prof. Peter McDonald Supervisor: Prof. Naohiro Ogawa Adviser: Prof. Bruce Chapman Abstract Recently, population ageing has become a major concern in developed countries. Of particular interest are the implications of ageing for labour markets and relevant policy options. This study focuses on one aspect of limited substitutability of different age-sex groups, as it presents a crucial issue in the policy debate on overcoming labour shortages in ageing economies. It concentrates on examining the relationship between changes in the age structure of the population aged + and employment of particular age-sex groups within labour force as it ages. The researched question is addressed by means of one and two factor decomposition of employment rate. The data was taken from Japanese and Australian censuses between 96 and.the decomposition considered age composition and employment rates as components of change in the overall employment rate across an intercensal period. Given age-sex specific employment patterns of particular industries, with the advances in the ageing of the labour force: - population structure becomes less favorable to the total employment - some industries benefits from the changes in the age composition, for most the age compositional changes tend to be unfavorable. - the benefits of debated policies are compromised by age-gender substitutability issues. Therefore they present only partial solutions - under present industrial structure and in the absence of significant productivity increases, the ageing labour force points towards labour force shortages.

3 LABOUR FORCE AGEING: ITS IMPACT ON EMPLOYMENT LEVEL AND STRUCTURE. THE CASES FROM JAPAN AND AUSTRALIA... Abstract... INTRODUCTION: AGEING OF LABOUR SUPPLIES DECELERATION OF GROWTH AND CHANGING AGE STRUCTURE OF LABOUR SUPPLIES LABOUR MARKET CONSEQUENCES AND RELEVANT POLICY OPTIONS....3 RESEARCH DESIGN... AGEING OF EMPLOYED AGEING OF TOTAL EMPLOYED AGEING OF EMPLOYED IN PARTICULAR INDUSTRIES CONTRIBUTING FACTORS TO AGEING OF EMPLOYED AGEING OF POPULATION AGED PROPENSITIES TO BE EMPLOYED Total employment Employment by industry CONTRIBUTION OF DEMOGRAPHIC AND ECONOMIC FACTORS AT DIFFERENT STAGES OF AGEING EFFECTS OF CHANGING AGE COMPOSITION AND EMPLOYMENT RATES ON TOTAL EMPLOYMENT EFFECTS OF CHANGING AGE COMPOSITION AND EMPLOYMENT RATES ON EMPLOYMENT IN PARTICULAR INDUSTRIES Tertiary industries Secondary industries Primary industries CONCLUDING REMARKS... 6 FUTURE POPULATION TRENDS AND POLICY CHALLENGES FURTHER ADVANCES IN THE AGEING OF POPULATION AGED INCREASING SIGNIFICANCE OF THE AGE-COMPOSITION COMPONENT Total employment Employment by industry POLICY CHALLENGES: INCREASING EMPLOYMENT OF AGED REFERENCES: APPENDIX... 37

4 Introduction: Ageing of labour supplies A prominent feature of dynamics of labour supplies in many developed countries over last decades has been deceleration of its growth and a changing age composition, skewing towards older age goups. Japan, having the oldest population in the world is the world forerunner in this process. By comparison, at present, the growth and age compositional changes of labour supply in Australia are only mildly advanced. Nonetheless if the current trends in fertility continue, and no measures are undertaken, in three coming decades Australia will have to face challenges of rapidly declining and ageing labour supplies that Japan is struggling now with.. Deceleration of growth and changing age structure of labour supplies The Japanese census estimated that the total labour supply at 66. million. Although this number was million higher than in 96, it was nearly million lower than in 99. This was the first time in the last 4 years that the size of the total labour force decreased. The decline of total supply of labour, observed in the period 99- was the result of stagnation and deceleration of the labour force growth, noticeable in Japan since 96 s. The labour force growth rate has declined from per cent per annum between 96-97, to. per cent between and declined further still to take negative values (-.3 per cent per annum) between 99 and (see: Figure ). Figure The size (in ) and growth (in percent per annum) of the total labour force: Japan 96-; Australia 97- per annum Japan Year (') per annum Australia Year (') Growth over preceding census year (left axis) Total size (right axis) Source: Author s calculations based on data from: Japan: Bureau of Statistics, Management and Coordination Agency, Population Census of Japan (Various years); Australia: Australian Bureau of Statistics, Population Census, (Various years) 3

5 The situation was different in Australia with regard to both: the size and dynamics of labour supply growth. According to census data, the total supply of labour in Australia was nearly 9 million, 7. times smaller than the labour supply of Japan. This number was by 6.6 per cent higher than in 996 (an increase of thousand) and 68 per cent higher compared to its size in 97 (an increase of 3.6 million). These increases are part of the high labour force growth rate observed in Australia over the last 3 years. The most rapid increase in the labour force occurred between 97 and 98 when the growth rate peaked at. per cent. Since then, the increases have generally slowed. In the most recent period (996-) labour supply growth was at.3 per cent per annum level. Although decline of the labour force share of aged -4 has occurred in both countries, their proportion was much lower in Japan at the end of researched period (.8 per cent in ), than in Australia (8 per cent in ). Similarly, the proportion of aged -4 started to decline in both countries, but at different dates. In Japan the declining trend was noticeable from 98 on, in Australia it has started only in 996. As result, the proportion of aged -4 was at the end of researched period much lower in Japan (6. per cent in ), than in Australia (7. in ). Similary, the share of aged + in the labour force started to increase in Japan from 96 on, in Australia it was noticeable only from 996 on. Consequently, while in Japan in the share of aged + (4 per cent) was more of the double of that aged -4 (.8 per cent), in Australia it consisted only 64 per cent of its share. Not surprisingly, the increase in medium age was higher in Japan than in Australia. While in Japan the median age has increased by nearly years from 34.3 years in 96 up to 43.9 years in. In Australia, the median age of labour force has risen up to 38. years in after having declined down to 9. years in 98 (see: Figure ). Figure The share of aged -4, -4, + in the labour force (in per cents) and the median age of the labour force (in years). Japan 96-, Australia Japan Year Years 6 4 Share of aged -4 in the total labour force (left axis) Share of aged -4 in the total labour force (left axis) Share of aged + in the total labour force (left axis) Median age of the total labour force (right axis) Australia Year Years 4

6 Source: As for Figure. Labour market consequences and relevant policy options Deceleration of the growth and subsequently decline of labour supply, accompanied by ageing described in the section above has serious economic implications. Without labour productivity gains, and given constant rate of technological progress and capital formation, the net result of ageing may be a decline of total economic output, and gradual economic slow down of the developed world. As future levels of productivity, technological progress and capital formation remain uncertain a number of policy options, aiming at stopping the potential and available labour supply from decline are being discussed. The most often quoted policy option is optimising the use of certain labour force reserves lying in low female and older persons participation rates and high unemployment rates. The gap between the existing employment compared to the state of the full employment and maximum labour force participation is quite wide for most developed economies. In view of this, a number of measures aiming at increasing employment, and rise participation of female and older workers is being proposed (Fotakis ; Hoehn ; Access Economics ; Ogawa 3; The Treasury 4). The underlying assumption of this policy debate is perfect substitution of between different age-sex groups within the labour force. That is not the case in the reality. A number of economic studies have proven that particular age-sex labour categories possess different skill-wage characteristics, which make them imperfect substitutes in the production process. In other words, decreases in relative supply of one category cannot be easily mitigated by increase in another without changes in their relative wages. We can expect that with rising educational levels of the younger labour force, the degree of the substitutability of the labour force will further decline, as the skill level embodied in a group of labour is a major determinant of its members substitutability for other workers (Easterlin 978; Freeman 979; Welch 979; Berger 98; Martin and Ogawa 988; Hamermesh ). Limited substitutability of these labour force groups is displayed in their different employment patterns, and distribution among industries. Any changes in the age-sex composition of the labour force will have impact on the employment patterns, and employment distribution among industries (Martin and Ogawa 984)..3 Research design Research question The author considers the above assumption of perfect substitution to be the major limitation of the debate. This study focuses on examining one aspect of the limited substitutability of particular age-sex groups within the ageing labour force. It concentrates on the relationship between changes in the age structure of the population aged years and over and employment of particular age-sex groups within the labour force as it ages. Comparing two countries (Japan-Australia) in different stages of ageing allows the simultaneous analysis of different levels of ageing. For further discussion on future levels of technological progress and capital formation associated with the ageing labour supplies see:(orzechowska )

7 Methodology and data The analysis concentrates on the comparison and decomposition of historical employment census data of two countries at different stages of ageing. The employment measure used in the analysis is the employment to population ratio (i.e. crude employment rate). The employment to population ratio stands for the number of employed persons (E..) as a percent of the population aged + (N..) and is calculated as follows: T.. = In order to show the effect of the changes in the age composition of the working age population on the employment, decomposition of the difference between two rates has been employed. The decomposition technique was developed by Kitagawa (9), slightly amended by Das Gupta (978) and widely applied by the number of others to the analysis of the age structural effect on employment patterns (Martin and Ogawa 984), occupational sex segregation (Bianchi and Rytina 986) as well as fertility, mortality and mobility rate differentials (Cho and Retherford 973; Wilson 988; Liao 989). In the analysis below, the difference between employment rates of two intercensal periods was decomposed into two major components: one attributable to changes in the demographic determinants of the available labour supply (the size and age composition of the population aged +), and the other due to changes in propensities to be employed. This has been done using the equation: ni ni Ni t. T. = (ti Ti) + Ti i n. i n. N. where: E.. N.. t, and T, total employment rates ti and Ti, employment rates of the ith age group; ni Ni, and shares of the ith age group in the population aged + n. N. ni (ti Ti) employment rate component: difference between age-standardized i n. rates ni Ni Ti age composition component: difference between employment ratestandarised age i n. N. compositions The major assumption underlying the technique is that the changes in the crude employment rates are only due to changes in the employment rates schedule and the age composition of population aged +. There are no residual terms. Another assumption is that there is no interaction between the two contributing components i.e. the age composition and employment rate components. For this reason, the major focus was taken on estimation of the significance and magnitude rather than causality of demographic and economic components lying behind changing levels of employment on the aggregate and industrial level. This technique also assumed that the changes in age composition and in specific employment rates occur linearly over the time. For this reason, the value of the technique is limited when applied to longer term analysis. The study presented stretches 6

8 over -year period of census data. It covers 4 years for Japan (96-) and 3 for Australia (97-). Outline of the paper After short introduction, the paper continue with section demonstrating ageing of employed as the core of the labour force ageing. Section 3 presents characteristics of contributing factors to ageing of employed Following this, the results of the decomposition analysis are presented in the section 4. Future population developments, its impact on employment and relevant policy options are the focus of section. Ageing of employed The ageing of employed lies at the core of aging labour supplies. Responsible for its ageing declines in the share of those aged -4 and increases in the share of those aged + in total labour supply are mirrored in the total employment.. Ageing of total employed The ageing of the employed labour force is slightly more advanced that the ageing of the total labour force. The increases of median age of employed were slightly higher than corresponding increases in the total labour force. As shown in Figure 3, in Japan the median age of employed increased from 34.3 years in 96 up to 44. years in, experiencing. years increase above that of the total labour force. In Australia the corresponding figures were 3.9 years in 97 and 38.9 years in, and the median age increase for this period was by.3 years above that of the total labour supply. Higher than of the total labour force, the median age of total employed, was due to high unemployment rates in youngest part of the labour force aged -4, resulting in their lower share in the total employment prevalent in both countries. 7

9 Figure 3 The share of aged -4, -4, + in total employment (in per cent) and median age of employed (in years) Japan 96- Australia Source: As for Figure Year 4 4 Years (left axis) -4 (left axis) + (left axis) Medium age (right axis) Year 996. Ageing of employed in particular industries The pace of ageing has been at different pace if looked at employment within particular industries. The highest median age of workers has been for primary industries. In Japan it has increased from 4.6 in 96 up to 63.4 years in. In Australia, the ageing of labour force in primary industries has been less rapid, its median age having risen from 4.6 in 97 up to 4. years in (see: Figure 4). Years 8

10 Figure 4 Median age of employed in primary industries (in years). Japan 96-, Australia 97- Years Japan Year Years Australia Year Source: As for Figure A. Agricultre, Forestry, Fisheries and Hunting Total primary industries Secondary industries, when compared to primary have aged at moderate pace in both countries. In Japan, their median age has increased by nearly years, from 3 years in 96 up to 44.9 in. In the oldest workforce of any secondary industry was in Mining with a median age of 48.8 years. Figure Median age of employed in secondary industries (in years). Japan 96-, Australia 97- Japan 96- Australia Years Year Years Year B. Mining C. Manufacturing E. Construction Total secondary industries 9

11 Source: As for Figure In Australia, the median age of workforce in secondary industries has risen less rapidly than in Japan, from 36. years in 97 up to 39. years in. As in Japan the highest median age of 39.8 years has been observed for workers in the Mining industries( figure). In contrast to primary, the tertiary industries had employed the labour force with the lowest median age, which nonetheless kept on rising. In Japan it has been on rise over the whole researched period. In Australia it has followed similar trend albeit with a slight drop in 98 (see: Figure 6). In Japan it has increased from 33 years in 96 up to 4. years in. In Australia, where, the median age increases were more limited than in Japan, it has risen from 34.6 years in 97 up to 38. in. The youngest labour force has been employed in: Electricity, Gas and Water Supply (median age of 4. years in ) and Services (median age of 4.8 years in ) in Japan, and Wholesale and Trade and Finance (median age of 34. years in ), Insurance, Property and Business (median age of 38 years in ) industries in Australia. Figure 6 Median age of employed in tertiary industries (in years). Japan 96-, Australia Years Source: As for Figure Japan Year Years D. Electricity, Gas, and Water Supply F. Wholesale and Retail Trade G. Transport, Storage and Communication H. Finance, Insurance, Property and Business I. Government and Defence J. Services Total tertiary industries Australia Year

12 3 Contributing factors to ageing of employed Ageing of employed was driven by both demographic and economic forces, which acted simultaneously, and together shaped the employment picture in Japan and Australia over the last 4, 3 years. 3. Ageing of population aged + The ageing of the population aged +, which consists the potential labour supply, was the major factor contributing to ageing of the labour force. Sustained fertility decline observed in recent decades, has translated into decline in the youngest group of the working age population, aged -4 in both Japan and Australia. In Japan the share of this age group has declined from 7 per cent in 96 down to 4.7 per cent in. Australia has seen a corresponding change. The proportion decreased from 4. 4 per cent in 97 down to 7. per cent in, having experienced a significant decline of 7. percentage points (see: Figure 7). Figure 7 The share of aged -4, -4, and + in population aged + (in per cent): Japan 96- and Australia Japan Year Years Australia Year Years Source: As for Figure Median age of population aged + (right axis) Share of aged -4 in the population aged + (left axis) Share of ged + in the population aged + (left axis) Share of aged -4 in the population aged + (left axis) The decline in the -4 age group has been accompanied by increases in share of the population aged +. This was especially rapid in Japan. The proportion of these aged + to those aged + has increased by 7.3 percentage points from 8. per cent in 96 up to 3.6 per cent in. In Australia it has increased between up to.6 per cent, experienced a slight drop in 99 down to.3, and subsequent increase up to 7.7 per cent in. Over the whole period it has remained on well-below- Japanese level. In Japan, the rapid increase of a number of persons aged + has been accompanied by decline of the share of the population in the prime working ageing (aged -4) in the

13 population aged +. This trend has been noticeable from 98 on. The proportion of aged -4 amongst these aged + from its peak of 8.8 per cent in 98, has been continuingly declining down to 49.8 per cent in. In contrast to Japan, Australia has seen significant increases of the proportion of aged -4 in the + population over most of 97- period. Its share has increased from.7 in 97 to.4 per cent in 996. Only recently in, the proportion of aged -4 has slightly declined down to. per cent. The above described age compositional changes(ie. rising proportion of those aged +, continuing decline of those aged - and -4) have resulted in rapid ageing of the population aged + in Japan and Australia. Not surprisingly, the process of ageing has been more rapid in Japan, than in Australia. Driven by extremely steep drop in the post-war fertility, the age compositional changes were very dramatic in Japan. As shown in the Figure 7, the median age of the Japanese population aged + has increased from 34.6 years in 96 up to 47. years in. In Australia, the median age of population aged + after a slight decline between 97 and 98 to 37.8 years, has been continuously rising to reach 4.3 years in 3. Propensities to be employed The age compositional changes in the population aged + translate into age compositional changes of those employed subject to age-sex specific employment rates (i.e. propensities to be employed or employment to population ratios). Therefore any changes in the age-sex specific pattern of employment have significant impact on the total level and age composition of employed. 3.. Total employment Over the last decades male and female employment level has been fluctuating in both countries. Both countries have witnessed decline of the male employment rates, visible on the Figure 8 as downwards shifts of the male employment rate curve for the following census years. In Japan the total male employment rate has declined from 84.4 percent in 96 down to 7.9 per cent in. In Australia it has decreased from 79. down to 6.6 per cent. The driving force behind the decline of male employment rates in both countries was a rapid reductions in the employment of aged + and -4. The employment rates of the oldest have declined by 7.8 percentage points over 96- in Japan and by 3.4 percentage points over 97- in Australia. Similarly, the employment rates of the youngest part of the male labour force aged -4, have decreased by.8 percentage points in Japan and 6. percentage points in Australia. The decline of male employment rates has been accompanied by initial decline (96-98) and subsequent increase of female employment rates in Japan and their steady increase in Australia (see upwards shifting curves in Figure 8) In Japan, over the whole 96-, the employment rates of particular age groups have followed opposite trends. Strong decline of -6.9 percentage points, observed for employment rates of females aged -4, has been accompanied by rapid rise of. percentage points for aged -4 and slight decline of percentage points for aged +.

14 In Australia, the increase in female employment rates has been observed for all age groups. The employment in the prime working age groups, was the strongest, their employment rate rose from 4. in 97 up to 6.3 per cent in (an increase of 4 percentage points). These aged + increased by 3.8 percentage points and aged -4 increased by.9 percentage points. Figure 8 Male and female propensities to be employed by age: Japan 96-, Australia 97- Japan 96- Australia 97- Male employment rates Age Female employment rates Age Source: As for Figure Age Age Although over the last four decades the levels of male and female employment rates have significantly changed, the age-sex specific pattern of employment in both countries remained virtually the same. As Figure 8 shows, in Japan and Australia male employment rates followed characteristic uni-modal pattern, with age of entry compromising the groups up to 9-4 and, prime working ages consisting of age groups between -4 years, and the ages of retirement, where employment rate drops considerably below its maximum begin with the group of to 9 years. Female employment rates over last decades have displayed distinct bi-modal pattern with two peaks around ages -4 and 4-49 separated by a through of lower participation rates in the central age groups (see: Figure 8). This characteristic bi-modally shaped pattern of the female labour force participation reflects Japanese and Australian females leaving the labour force after marriage or birth of children. In both countries females are more likely to work as part-time or temporary workers, and therefore subject to greater probability of lay-offs as companies attempt to adjust their labour force 3

15 3.. Employment by industry Changes in the level of the total employment have been accompanied by ongoing structural change of employment in Japan and Australia. This structural change has been reflected in changing employment growth and share, as well as in changing employment to population ratios of particular industries, i.e. propensities of the population age + to be employed in these industries. Tertiary industries With regard to changing propensities to be employed, the distinctive future of the Australian and Japanese structural change over last 3 years was continuously growing importance for employment of tertiary industries. Increasing level of employment in these industries is visible on the Figure 9 as upwards shifting employment rate curves for the following census years. In Japan the employment rate of tertiary sector has increased from.4 per cent in 96 up to 38 per cent in, having a total rise of.6 percentage points. Such an increase was caused mostly by rapidly expanding employment in the Service industries, whose employment rate grew by 8.3 percentage points between 96-. The second tertiary industry, with a rapid growth in employment has been the Wholesale and Retail. Its employment rate, although slightly declining from 98, when compared to its 96 level has increased by 3 percentage points. Employment in Australian tertiary industries has undergone an increase of 9 percentage points from 3.3 per cent in 97 up to 4.3 per cent in. This increase was mostly due to rapidly growing employment rate of Services (a total increase of 6. percentage points) and Finance, Insurance and Real Estate industries (a total increase of 4.3 percentage points). The employment rate in the tertiary sector has been growing due to both increasing employment of both: males and females. However, the increases for females were more significant. While the employment rates of males in the tertiary sector has increased over the whole 96- by 9. percentage points, the corresponding increase for females was.7 percentage points. The causes of rapidly growing female employment are associated a special demand in these industries for part-time workers during peak service hours, and partly women s desire for jobs that minimize conflict with child rearing and other household responsibilities (Retherford, Ogawa et al. 996). Similarly to Japan, in Australia, the rise of the total employment rate in tertiary industries was a great extent due to increase of female rather than male employment rates. Between 97- female employment rate has experienced a total growth of 3. percentage points while the corresponding increase in the male employment rates was only. percentage points. As in Japan, the Australian female employment rose most rapidly in the Services, where the female employment rates have increased by 8.9 percentage points from per cent in 97 up to 9.9 per cent in. 4

16 Figure 9 Male and female age-sex specific propensities to be employed in tertiary industries: Japan 96-, Australia 97- Japan 96- Australia 97- Tertiary industries, male employment rates Age Tertiary industries, female employment rates Age 4 3 Source: As for Figure Age Age Regardless changing level of male and female employment rates in tertiary industries, the age pattern of employment in these industries remained the same over the whole researched period in both countries (see: Figure 8). Tertiary industries in Japan displayed a strong tendency to employ younger labour force. The peak of their male employment was reached between -39 age groups, owing to high employment rates of these age groups in the Wholesale, Retail and Trade, Finance, Insurance, Property and Business and Services (see: Figure in Appendix). The female employment displayed bimodal distribution with two peaks at ages -4 and 4-44, nonetheless the employment peak at younger ages was much higher. This was because a number of tertiary industries such as: Electricity, Gas and Water Supply, Finance, Insurance, Property and Business, Government and Defence have displayed have only one age group -4 of maximum employment (see: Figure 6 in Appendix.). In Australia, unlike in Japan, the highest male employment rates in tertiary industries on the whole have been observed in the in prime working age groups Nonetheless some tertiary industries such as Wholesale Retail and Trade, and to lesser extent Finance, Insurance, Property and Business have the highest employment rates in younger age groups. The Wholesale, Retail and Trade had the highest likelihood to employ very young males ages -4, and the Finance, Insurance and Property young males aged -39 (see: Figure in

17 Appendix). Female employment rates in tertiary industries had two peaks at younger ( - 4) and middle age groups (4-49). Both of them were at similar level. Notwithstanding this, when looking at the age pattern of female employment in particular tertiary industries it can be easily noticed, with exception of services, they have only one peak of female employment at younger ages (see: Figure 6 in Appendix.). Secondary industries Unlike that of tertiary, the employment rates secondary industries have been declining in both countries. This is visible on the Figure as downwards shifts of employment rate curves for the following census years. In Japan, after slight rise up to.6 per cent in 97, the employment rates of secondary industries have gone through a rapid decline of. percentage points down to 7. per cent in. This decline was driven by drop of the employment rate in the Manufacturing industries, which between 97- decreased by.8 percentage points. Declining employment in Manufacturing has been accompanied by a slight decrease of employment rates in Mining (-.7 percentage points over 96-), and a relatively stable -6 percent level in Construction. The decline of total employment rates in secondary industries was owing to decline in both male and female employment, which has decreased by 6. and 3.8 percentage points respectively. Similar trends could be found in Australia. The secondary industries went through employment decline over the 97-. Their employment rates have decreased from 8.8 per cent in 97 down to. per cent, having experienced a decline of 7.7 percentage points. Manufacturing was the hardest hit by this decline in employment. While in 97 it had the highest employment rate of all industries equal to 3.4 per cent, in it was only 6.8 per cent.the decline in the total employment in secondary industries was stronger among males than females. While male employment rates have decreased by.3 percentages points over the 97- period, the corresponding decline in female employment was just.8 percentage points only. 6

18 Figure Male and female age-sex specific propensities to be employed in tertiary industries: Japan 96-, Australia 97- Japan 96- Australia 97- Secondary industries, male employment rates Age Secondary industries, female employment rates Age cent Per Age Age Source: As for Figure As in the case of tertiary industries, the age-sex pattern of employment remained quite stable over the years stable, regardless of changing employment level (see: Figure ). In both countries it has closely reflected the aggregate employment pattern with highest male employment rates observed in the prime working ages (-4). Female employment rates for the secondary industries took on the whole bimodal distribution with two employment peaks at younger and middle ages. Notwithstanding this, some industries such as Mining and Construction in Australia and Mining in Japan had only one peak in younger ages (see: Figure 8 in Appendix). Primary industries Primary industries in Japan and Australia have experienced the steepest decline of the employment rates among all the industries (see on Figure downwards shifting employment rate curves for the following census years). In Japan, over the 96- period, they have undergone a total decrease of 9 percentage points from per cent in 96 down to 3 percent in. The decline of employment was of similar magnitude for male and female workers, whose employment rates have declined by 8.4 and 9. percentage points respectively. This rapid decrease of employment rates in the primary sector was the result of structural changes in the Japanese economy that have caused a 7

19 substantial reallocation of workers from self-employment and family work in the agricultural sector towards employment in other sectors of the economy (Furuya and Clark 993). In Australia, like in Japan, the employment rate of primary industries has been decreasing. Compared to Japan though, the drop of employment rates was of limited magnitude, as they decreased by only per cent from 4.3 per cent in 97 down to. per cent in 97. Figure Male and female age-sex specific propensities to be employed in primary industries: Japan 96-, Australia 97- Japan 96- Australia 97- Primary industries, male employment rates Age Primary industries, female employment rates Age Source: As for Figure Age Age Disregarding changes in employment level of primary industries, the age employment pattern remained the same over the last decades in both Japan and Australia. As Figure shows, over the whole researched period primary industries have tended to employ adult (aged -4) and older (aged +) labour force. 4 Contribution of demographic and economic factors at different stages of ageing The demographic and economic factors contributing to ageing, described in above section, acted on employment simultaneously. The section below presents the results of the decomposition analysis, which distils ageing into demographic and economic factors. These are quantified into two components: age composition and employment rate components. 8

20 The age composition component represents changes in the employment rate between two periods that would have occurred due to changes in the age composition of the population aged +, given no changes of employment demand take place. The employment rate component, on the other hand, presents changes in the employment rates of two periods given changes in the employment demand, assuming no fluctuations in the age composition of the population aged + 4. Effects of changing age composition and employment rates on total employment The magnitude and direction of age composition and employment rate components have varied over decades corresponding to the changing in demographic and economic factors in both countries. In Japan, the ageing of the population aged years and over last four decades, resulted in an increasingly negative age composition component (see: Figure ). During the first two decades of the researched period (96-98), when the ageing of the population was only mildly advanced (i.e. growth of the aged +, and decline of -4 have been accompanied by still increasing share of the aged -4 see: Figure 7, page), the age composition component took positive values, being.4 percentage points between 96 and 97 and.8 between 97 and 98. The last two decades (98-), in which the process of population ageing have accelerated, have witnessed negative values of the age composition component. Between 98 and 99, when the share of aged -4 in the + population started rapidly rapidly declining, and the proportion aged + rapidly increasing, the age composition component dropped down to -.7 level. It has remained negative, but lower (-.36 per cent) level over 99-, when the process of ageing slowed down slightly. As shown in Figure, the magnitude of the age composition component was far more limited in Australia. Between 97and 98, when growth of aged -4 age groups stagnated and the -4 age group started declining, the age composition component took negative values (-.74 per cent). In the following years, with acceleration of the growth of -4 age group and its increasing share in the population aged +, the value of the age composition component have increased up to -.34 between 98 and 99 and.7 per cent between 99 and 9

21 Figure Age composition and employment rate components for the total employed: Japan 96-, Australia 97- Japan 96- Australia year intervals year intervals Source: As for Figure Age composition component Employment rate component The values of the employment rate component have varied for both countries in different periods, reflecting fluctuations in the employment demand of economies. In Japan, the employment rate component took negative values for most of researched period. Its extreme low value (-3.8 per cent) between 97-8 can be associated with rapid rise of unemployment in Japan in that period. The unemployment rate has increased from.3 per cent in 97 up to.48 per cent in 98. Similarly the positive value of employment rate component (.4 per cent) between 98-9 can be correlated with a drop in the unemployment rate, observed in 9 s. Other than in Japan in Australia the employment rate component was positive for most of researched period. Its negative values between 98-9 (-.6) are to be associated with economic recession and rising unemployment rates. The unemployment rates have increased from.9 per cent in 98 up to.6 per cent in 99. Author s estimates of unemployment rate based on data from: Japan: Bureau of Statistics, Management and Coordination Agency, Population Census of Japan (Various years); Australia: Australian Bureau of Statistics, Population Census, (Various years)

22 4. Effects of changing age composition and employment rates on employment in particular industries The age composition and employment rates components have varied widely in their value and direction for employment in particular industries. While for employment in some industries the age compositional changes in the population aged + were beneficial for others they were unfavorable (i.e. the age compositional component took negative values). The negative age composition component for the total employed, prevalent in Japan over last 3 decades and in Australia between 97 and 99, was mostly due to its negative effect on employment in tertiary and secondary industries (see: Figure 3). Primary industries in both countries, on the other hand, largely benefited from age compositional changes. Figure 3 Contribution of particular industries to age composition component: Japan 96-, Australia Japan Australia Source: As for Figure year intervals -. Tertiary Primary Secondary year intervals The value and direction of the employment rate component has mirrored the direction of structural change of employment in Japan and Australia. Corresponding to tertiary industries increasing share in employment, employment rate component in these industries remained positive in Japan and in Australia (see: Figure 4). In contrast, primary and secondary industries, whose share in the total employment has been at decline in both countries, displayed for most of researched period highly negative values of employment rate component.

23 Figure 4 Contribution of particular industries to the employment rate component: Japan 96-, Australia Japan Source: As for Figure year intervals - Primary Secondary Tertiary Australia year intervals 4.. Tertiary industries It is not surprising that recent age compositional changes of population aged + were not favourable towards employment in secondary and tertiary as industries. These industries employed the highest proportion of young (-4) and adult (-4) workers (see: section 3.., page 4). The age compositional component took the strongest negative values for tertiary industries. In Japan, it has been down to -.7 per cent between 98 and 99, and slightly rising in up -.9 per cent between 99 and, corresponding to the changing share of people aged -4 and -4. In Australia, the lowest, calculated, value of the age compositional component has been for the 97-8 period, when the decline in the -4 age group was accompanied by stagnation of -4 age group. Since then the value of the age composition component has been increasing. This is mainly the result of continuously growing -4 age group in the population aged +, its growth compensating for the decline of the -4 year olds (see: Figure 3). Figure shows, that among all tertiary industries, the age composition component has taken the lowest negative values for Services in Japan and Wholesale, Retail and Trade in both countries. These were the industries of the highest employment share of the youngest age groups (see: Figure and Figure 6 in Appendix). The negative age composition component was also of significant magnitude for Transport, Storage and Communication in Japan and Finance, Insurance and Property and Business in both countries.

24 Figure Age composition component for particular tertiary industries: Japan 96-, Australia 97- Japan 96- Australia Source: As for Figure year intervals -.3 D. Electricity, Gas and Water Supply F. Wholesale, Retail and Trade G. Transport, Storage and Communication H. Finance, Insurance, Property and Business I. Goverment and Defence J. Services year intervals The age composition component, has for most tertiary industries acted in the opposite direction to the employment rate component, which has been positive for most tertiary industries (see: Figure 6). The increasing employment demand and share of Services in both Japan and Australia and Finance, Insurance, Property and Business and Wholesale and Retail in Australia have been reflected in their increasingly positive employment rate component. 3

25 Figure 6 Employment rate component for particular tertiary industries: Japan 96-, Australia 97- Japan 96- Australia Source: As for Figure year intervals - D. Electricity, Gas and Water Supply F. Wholesale, Retail and Trade G. Transport, Storage and Communication H. Finance, Insurance, Property and Business I. Goverment and Defence J. Services year intervals 4.. Secondary industries The negative value of the age compositional component for the secondary industries was smaller than for tertiary. This is because they have employed the highest proportion of adult labour force, the number of young workers, aged -4 employed in these industries was smaller than in tertiary sector. As such secondary industries were the most sensitive to the changes the share of 4-4 age group in the population aged +. Decline of this age group in the population aged +, observed in Japan from 98, and its stagnation between 97-8 in Australia, has translated into negative values of age composition component for these periods in both countries. In Japan, its value between 98- oscillated between -.84 and -.6 per cent. In Australia, it has dropped to -.3 per cent between 97-8, which following started to increase in its value, correspondingly to the changes in the share of aged -4 and -4 in the population aged + (compare: Figure 7 and Figure 3). In both Japan and Australia the strongest, negatively affected has been employment in Manufacturing, and Construction (see: Figure 8). Unlike for tertiary industries, the negative age composition effect in secondary industries has been accompanied by negative employment rate component in both countries (see: Figure 4, page: ). As shown on Figure 7, Manufacturing industries made a major contribution the negative employment rate component of secondary industries in both Japan and Australia. 4

26 Figure 8 Age composition component for particular secondary industries: Japan 96-, Australia 97- Japan 96- Australia Source: As for Figure year intervals -. B. Mining C. Manufacturing E. Construction year intervals Figure 9 Employment rate component for particular secondary industries: Japan 96-, Australia 97- Japan 96- Australia Source: As for Figure year intervals -3 B. Mining C. Manufacturing E. Construction year intervals

27 4..3 Primary industries Primary industries, benefited the most from the recent age compositional changes in both countries. This is not surprising as these industries, as shown in section 3.., displayed the strongest tendency to employ older workforce over last 4, 3 years. With increasing share of older age groups in the population aged +, the value of the age composition component took positive values in both countries. It has oscillated between.4 and.94 per cent in Japan and between.4 and.9 per cent in Australia (see: Figure 3, page: ). The magnitude of the age compositional component for Australian primary industries has been far more smaller than for Japanese, due much lower proportion of aged + in the population aged +. Changes in the age composition of the population aged + favourable towards employment in primary industries did not correspond with the direction of the structural change of employment. The positive age composition component has been accompanied by negative employment rate component over the whole researched period in both countries. As depicted in Figure 4, (page ) the value of the employment rate component in Japan has been the lowest between 96-7 (-9.66 per cent), when the biggest decline in the employment share of primary industries has occurred. Since then, with deceleration of decline in the employment in primary industries, the employment rate component has been increasing in value, although remaining negative. In Australia, between 97 and the employment rate component oscillated between -.98 (98-9) and.7 per cent (99- ). 4.3 Concluding remarks The decomposition of the difference between employment rates of two periods have shown that: - Given stable age-specific employment patterns, ageing of the labour force structure became less favourable towards total employment. This has been shown in the increasingly negative age composition component grows, as the labour force ages. While decline of the -4 and increase in the aged + alone resulted in moderate changes in the significance of the age composition component, accompanied by decline of the prime working age group (aged - 4) brought along its strong negative values. This is illustrated by comparison of values of age composition components in Australia and Japan over last decades. They were much lower for Japan where the rapid decline of the aged -4 started in 8 s. In Australia, where ageing of the labour force has been only visible in the decline of aged -4 and increasing proportion of aged +, the magnitude of the age compositional component has remained for the whole researched period on much lower level. - The significance and the direction of the age composition component have varied for different industries, depending directly on the age-pattern of their employment. 6