Australian Government, 2007–08 Budget

Statement 4: Australia's Labour Force Utilisation

This statement examines the progress that Australia has made on labour force utilisation, and considers the impact of the future demographic challenge. Labour force utilisation is one of two key policy priorities identified in the Intergenerational Report 2007, released in April 2007, for addressing the inevitable slowing of economic growth arising from demographic change. The other key priority is productivity growth.

The ongoing rise in employment is one of Australia's important economic and social success stories. The participation rate reached an all-time high in September 2006, at just over 65 per cent. Coupled with this, unemployment is currently at a 30-year low.

This statement begins with a profile of Australia's path of economic growth, contrasting this with other OECD countries in order to provide international context for Australia's position on the spectrum of labour market arrangements and outcomes.

The statement then turns to Australian labour market performance in an historical context. The growth in employment, decline in unemployment and the change in the composition of the labour force over the past decade and a half are remarkable. These have been important contributors to Australia's impressive record of economic growth. The labour market outcomes have been brought about through a mixture of policy and non‑policy factors: broad policies associated with macroeconomic stabilisation and creating the conditions for sustained growth; specific policies designed to provide incentives for greater participation in the labour force; and changing demographics, boosting the availability of people in working-age cohorts.

The statement then examines the growth of casual labour and scrutinises the perceived trade-off between flexible labour markets and job security. Finally, policy priorities for maintaining full employment in the face of global change are considered. While there remains further scope to improve labour market outcomes, progress becomes more difficult as Australia approaches global best practice. Moreover, at the limit, labour utilisation is bounded by the number of hours people want to work. As the Intergenerational Report projects, the proportion of the population made up of the working-age cohort is likely to dwindle primarily because of the retirement of the baby boomers.

The challenge for policy is to obtain the best labour force outcomes as the age structure of the population changes, while at the same time improving the productivity performance.

Labour force utilisation in international and historical context

In this section, Australia's labour market performance is discussed in international and historical context. Trends in labour utilisation are examined by considering developments in participation rates, demographics, unemployment, hours worked and labour mobility.

Labour utilisation and productivity in international context

GDP per capita is a standard measure of the performance of an economy over time and across countries. GDP per capita can be expressed as hours worked per capita multiplied by GDP per hour worked. The former is a measure of labour utilisation and the latter is the standard measure of labour productivity.

Chart 1: Path of economic growth — Australia

Chart 1: Path of economic growth — Australia

Source: Groningen Growth and Development Centre, Total Economy Database, January 2007.

The horizontal axis in Chart 1 shows labour utilisation measured as weekly hours worked per capita. It depends on demography (that is, the age structure of the population), labour force participation, unemployment, and average hours worked. The vertical axis shows labour productivity measured as GDP per hour worked. Growth in GDP per capita over time is shown by the north-east trend in the chart.

In 1960, the average Australian worked 15 hours a week and an average hour of work generated $22.90 of GDP, measured in 2005 dollars. Multiplying these two figures together, we find that, expressed in 2005 dollars, real GDP per capita was $343.50 a week. In 2006, the average Australian worked 17.1 hours a week and an average hour of work generated $56.77 of GDP. Hence, in 2006, real GDP per capita was (multiplying the two numbers together) $970.77 a week.

Subsequent to the last recession, average annual growth in GDP per capita has been a little under 2.5 per cent. Economic growth has been achieved by growth in both labour utilisation and productivity. The large shifts to the right following 1983 and 1993 are a consequence of falling unemployment associated with recovery from recession.

Chart 2 plots the same relationship, but across countries rather than over time. The curved lines running through Australia and the US represent combinations of labour utilisation and productivity that correspond with the same level of GDP per capita. For instance, Netherlands has around the same GDP per capita as Australia but achieves this with lower labour utilisation and higher productivity. Alternatively, a country could have similar GDP per capita but with a combination of higher utilisation and lower productivity. Countries above the line running through Australia have a higher level of GDP per capita.

Chart 2: Labour utilisation and productivity in OECD countries, 2006

Chart 2: Labour utilisation and productivity in OECD countries, 2006

Source: Groningen Growth and Development Centre, Total Economy Database, January 2007.

Labour utilisation and productivity make different contributions to generating GDP per capita, as can be seen by comparing Australia, the US and France. Australia and the US have very similar labour utilisation. The key difference between the two is the level of labour productivity, which is around 20 per cent higher in the US than in Australia. In contrast with Australia, France has significantly lower labour utilisation with labour productivity roughly similar to the US, leading to lower overall GDP per capita. There are several factors contributing to lower labour utilization in France and other European countries. Blanchard (2004) argues that European countries have used the increase in productivity over time to increase leisure rather than income, but others argue that labour utilisation outcomes are associated more with labour market and social policy settings.

Chart 3 shows that labour utilisation has been steadily declining in France since the 1960s. A succession of regulations since the 1980s together with demographic change have gradually reduced average weekly hours worked (OECD 2005). In 1998, the government mandated a reduction in the working week from 38 hours to 35 hours for large firms by 2000 and for smaller firms by 2002. Employment protection legislation in France has tended to inhibit not only the firing but also the hiring of workers.

In France, gains in real GDP per capita have been achieved by trend productivity growth more than off-setting the trend decline in labour utilisation, illustrated by a north-westerly progression in Chart 3. In contrast, in the United States, which is characterised by a considerably more flexible labour market, both productivity and labour utilisation have trended up over time, resulting in a north-easterly progression in Chart 3.

Chart 3: Labour utilisation and productivity

France

 

United States

Chart 3: Labour utilisation and productivity - France

 

Chart 3: Labour utilisation and productivity - United States

Source: Groningen Growth and Development Centre, Total Economy Database, January 2007.

Bourles and Cette (2006, 2005) find that some European countries have higher labour productivity because they have lower levels of employment and lower productivity workers are less likely to participate in the labour force. McGuckin and Van Ark (2005) find a negative relationship between labour utilisation and productivity, but only in the short run. This result suggests that new labour force entrants have lower productivity but catch up with the rest of the labour force after a period of skills development.

These comparisons highlight an important distinction between labour utilisation and labour productivity as drivers of income level. A given level of income can be reached by either higher labour utilisation or higher productivity (or a combination of both). However, labour utilisation is bounded by the number of hours people want to work. By contrast, productivity growth is not bounded (although the current productivity level is bounded by the technology frontier).

Decomposition of labour utilisation

In Australia, labour utilisation has increased over time, with average weekly hours per capita increasing from around 15.5 hours in the 1960s to around 17 hours in 2006. Changes in labour utilisation can be decomposed into changes in: labour force participation; population; unemployment; and hours worked by employed people. While growth in total population tends to be steady and has little impact on the swings in labour utilisation, the age distribution of the population has a large impact on labour utilisation.

All other things being equal, an increase in labour utilisation would result from an increase in the participation rate, an increase in the working age proportion of the population (15 to 64 year olds), an increase in average hours worked or a decrease in unemployment.

Participation

Chart 4 shows that the participation rate, the ratio of the labour force to the population aged 15 years and over, increased gradually from around 61 per cent in the late 1970s to 64.8 per cent in March 2007. An all time high was reached in September 2006.

Chart 4: Participation rates (total)

Chart 4: Participation rates (total)

Source: ABS cat. no. 6202.0.

Behind this development has been an increase in female labour force participation which has more than offset a decline in male participation (Chart 5). The female participation rate has increased from around 44 per cent in the late 1970s to 57.5 per cent in March 2007, lifting total participation over the same period. Similarly, participation rates for students also increased over this period.

Chart 5: Participation rates (male and female)

Chart 5: Participation rates (male and female)

Source: ABS cat. no. 6202.0.

Strong employment growth since 2004 has encouraged more people to enter and remain in the workforce. In particular, the participation rate of males appears to have ceased its long-term decline and has recently increased slightly.

Australia's participation rate exceeded the OECD average of 60.2 per cent in 2005 but was below the rates of nine other countries, including Canada, Denmark, Norway, Switzerland, Sweden and the US.

Abhayaratna and Lattimore (2006) identified the key cohorts in which Australia's participation is relatively low. For prime aged males (25 to 54 years), Australia ranked 25th among 30 OECD countries in 2005, more than 5 percentage points below the top ranking countries. For child-bearing aged females (25 to 44 years), Australia ranked 23rd, well behind the rates of New Zealand, the US, the UK and Canada. For people nearing retirement (55 to 64 years), Australia ranked 13th, behind the rates of New Zealand, the US, the UK and Canada. If Australia closed the participation gap with the highest ranking comparable OECD country in 2005 for each of these labour market segments, there would have been around an additional 600,000 people participating in the labour force.

Several countries with higher participation rates than Australia have much higher participation amongst older workers. For example, participation rates in 2005 amongst 55 to 64 year olds were 86 per cent in Iceland, 73 per cent in Sweden, 71 per cent in New Zealand, 69 per cent in Norway, 68 per cent in Switzerland, and 67 per cent in Japan. The same rate for Australia was around 55 per cent.

Nonetheless, participation amongst older workers in Australia has risen significantly over recent years. Kennedy and Da Costa (2006) find that the increase in the participation of women aged 55 to 64 years since the early 1990s is evenly spread between part-time and full-time participation. For men, full-time participation fell between the late 1970s and mid-1990s before rising sharply in 2000. Part-time participation has been on a steady upward trend over this whole period. Over time, different generations of men have not had markedly different participation rates in older years except the most recent generation, but more recent generations of women have had higher participation. This suggests that the upward trend in female participation will continue for some time yet. It is too early to tell whether the recent increase in older male participation will be sustained.

Kennedy and Da Costa also suggest that tightening eligibility for the Disability Support Pension, the mature age tax offset and the recent changes to superannuation might be part of the explanation for higher participation of older workers. Moreover, increasing educational attainment of younger generations may contribute over time to higher participation rates. There may be family reasons — increased participation by women might be causing higher participation by men, if partners choose to retire at the same time. Attitudes about hiring older worker could be changing. Older workers may be reducing their hours and working part time as a transition to retirement rather than giving up work altogether. It may also be a behavioural response to increasing life expectancy and concerns about retirement income adequacy. Finally, the cyclically strong labour market could be causing fewer separations of older workers and so reducing 'unplanned' retirement.

The Intergenerational Report 2007 projects that Australia's overall participation rate will fall from around 65 per cent to around 57 per cent by 2046-47 due to the ageing of the population. Notwithstanding that the participation rate of people of traditional working age (15 to 64 year olds) is expected to rise from 76.2 per cent in 2006-07 to 78.1 per cent over the same period the participation rates of older cohorts are also expected to rise, but not to the levels of younger cohorts. With increasing shares of the population in older cohorts, the overall participation rate will fall. These trends highlight the important role played by demography on participation.

Demography

One way of illustrating the importance of demography on labour utilisation is to examine the dependency ratio: the number of dependants — the young aged 0 to 14 and those aged 65 and older — divided by the number of working age population. A decline in the dependency ratio lifts labour utilisation in two ways. First, it means the proportion of working aged people in the population increases. Second, the albeit smaller effect, is that fewer dependants need fewer unpaid carers to look after them. As Chart 6 shows, the dependency ratio declines steadily from the 1960s as the sharp decline in young dependants more than offset the gradual rise in the proportion of older persons until we reach a demographic 'sweet spot' in around 2010. After this time the rise in older persons dominates and demography begins to detract from labour utilisation.

Chart 6: Dependency ratio — Australia

Chart 6: Dependency ratio — Australia

Source: United Nations 2006 Revision Population Database, medium variant.

In Australia, as with many developed countries, population ageing will reduce labour utilisation over the next 40 years. The Intergenerational Report 2007 projects that in the period to 2047 the number of 0 to 14 year olds in Australia will rise slightly, those of traditional working age (15 to 64 years) will rise by about 20 per cent, older people (65 to 84 years) will more than double and the very old (85 years and over) will more than quadruple. The dependency ratio (both young and old as a proportion of working age people) will rise from around 50 per cent in 2007 to nearly 70 per cent in 2047. On current participation rates by age cohort, this implies a significant reduction in the availability of labour.

Australia is by no means alone in facing a rising dependency ratio — in fact, Australia's demographic challenge in this respect is much less serious than in some other countries. The dependency ratio averaged 65 per cent in the OECD area in 2005 and ranged between about 55 per cent in Korea, the Czech and Slovak Republics and above 80 per cent in Mexico and Turkey. The dependency ratio is expected to rise sharply in most OECD countries, except Mexico and Turkey where it is expected to decline. By 2050, the dependency ratio is expected to exceed 100 per cent in Italy, Japan and Spain.

The rise in the dependency ratio comes about despite a fall in the proportion of young people. The youth‑dependency ratio fell from around 60 per cent in 1980 to around 40 per cent on average in the OECD area in 2005. In most OECD countries, the youth dependency ratio is projected to continue to decline but only modestly to average 37 per cent in 2050. On the other hand, the elderly-dependency ratio across the OECD has risen from around 20 per cent in 1980 to around 24 per cent in 2005. By 2050, this ratio is projected to more than double in the OECD area to 52 per cent. Some countries will face much sharper increases. By 2050, the old‑age‑dependency ratio will exceed 70 per cent in Italy, Spain and Japan (OECD 2006c).

The impact of demography can also be shown by disaggregating the increase in utilisation into demographic and non-demographic factors. The solid line in Chart 7 shows that labour utilisation has increased by around 1.2 hours per week since the late 1970s. The bars represent contributions to the change. The demographic effect has increasingly contributed to higher labour utilisation as working age cohorts have increased relative to the total population. From the turn of this decade, the demographic effects are projected to be negative (for both males and females). The non-demographic effects include the impact of business cycles on unemployment and average hours, plus changing participation rates and hours worked within age cohorts. Non-demographic effects have mostly added to labour utilisation only in recent years.

Chart 7: Contributions to the increase in labour utilisation

Chart 7: Contributions to the increase in labour utilisation

Contributions to the change in labour utilisation are calculated using a rolling annually re-weighted shift-share approach.

Source: Australian Bureau of Statistics and Treasury calculations.

The underlying trends are quite different for males and females. For males, the non-demographic effects have been negative, more than offsetting the demographic effects, resulting in average weekly hours worked per male falling by around an hour per week (Chart 8).

Chart 8: Contributions to the increase in labour utilisation (males)

Chart 8: Contributions to the increase in labour utilisation (males)

Source: Australian Bureau of Statistics and Treasury calculations.

For females, both demographic and non-demographic effects have contributed to increased average hours worked, with non-demographic effects having a larger influence than demographic effects. Average weekly hours per female have increased by over 3 hours per week (Chart 9).

Chart 9: Contributions to the increase in labour utilisation (females)

Chart 9: Contributions to the increase in labour utilisation (females)

Source: Australian Bureau of Statistics and Treasury calculations.

Average hours

While labour utilisation overall has increased, average weekly hours worked per employed person have fallen by over two hours since the late 1970s (Chart 10).

Chart 10: Average weekly hours of those in work

Chart 10: Average weekly hours of those in work

Source: ABS cat. no. 6291.0.55.001 and Treasury calculations.

The main source of the decline is an increase in the proportion of part-time workers (Chart 11). On average, part‑time employees work around 16 hours per week compared with around 40 hours for full-time employees.

Chart 11: Part-time share of total employment

Chart 11: Part-time share of total employment

Source: ABS cat. no. 6202.0.

Increased part-time work should be regarded as a positive development if it reflects individuals' preferences for labour and leisure. Casual employment is taken up in a latter part of the statement.

While average hours worked per employee have been falling as the share of part-time employment rises, it has been suggested that Australians in full-time employment have been working longer average hours than any other OECD country. However, at around 40 hours per week, average full-time hours are currently around the same level they were three decades ago. In addition, recent work by the Australian Bureau of Statistics (Baker and von Sanden 2006) suggests that there has been a consistent overestimate in the hours measures used to rank Australia against other OECD countries. Once this has been accounted for there is no evidence to suggest that Australians working full-time are working more, or indeed fewer, hours per week than workers in other OECD countries.

Unemployment

Increased participation alone will not lift labour utilisation unless those new participants are employed. The unemployment rate has fallen to a 30-year low (Chart 12). Following the oil shock in the early 1970s, the unemployment rate rose sharply from around 2 per cent to around 6 per cent between 1974 and 1978. This reversed a steady trend towards increasing labour utilisation underway since the 1960s. Following the recessions in the early 1980s and early 1990s, unemployment again rose sharply with the result that over 10 per cent of the labour force were without a job. Strong employment growth since the early to mid-1990s has resulted in a steady fall in the unemployment rate.

Chart 12: Unemployment rate

Chart 12: Unemployment rate

Source: ABS cat. no. 6202.0.

Australia's unemployment rate is below the OECD average, and lower than major economies such as the UK and Canada and similar to that in the US. Moreover, Australia's unemployment rate has fallen substantially more than the OECD average over the last decade and a half (Chart 13). On the OECD's standardised unemployment rate measure, the most recent peak in both the OECD average and Australian unemployment rates was in 1993, when the OECD average was 7.8 per cent and the Australian rate was 10.6 per cent. The gap between unemployment rates was virtually closed by the end of the last decade. In more recent years, Australia's unemployment rate is a full percentage point or more lower than the OECD average unemployment rate.

Chart 13: Difference between Australian and OECD unemployment rates

Chart 13: Difference between Australian and OECD unemployment rates

Source: OECD Main Economic Indicators database.

Reduction in the unemployment rate is likely to be assisted by a mobile labour force. This aspect is taken up in Box 1.

Box 1: Labour mobility

A flexible, mobile labour force plays an important role in maintaining macroeconomic stability by preventing wage pressures in one region or occupation spilling into more generalised wage pressures and inflation at the national level. Labour mobility also facilitates economic growth by enabling the reallocation of labour to its most productive use, including in order to exploit new opportunities and technologies as they arise.

In an economy operating at close to full employment, relative wage adjustments become even more important in attracting workers from other industries or regions. Under a rigid, centralised wage fixing system, wage dispersion would be low and firms in more productive sectors of the economy would find it more difficult to attract workers from less productive sectors. As wage setting becomes more decentralised and responsive to sectoral and regional conditions, the dispersion of wages is higher resulting in higher labour mobility, a lower dispersion of unemployment and improved national productivity.

Chart A shows the range of annual growth in average weekly earnings in the six States. The increasing volatility of this indicator suggests that wage flexibility across the States has been increasing. As discussed in Budget Statement 3 increased wage dispersion provides a clearer signal about changes in labour demand in industries and regions, and is one factor that will increase labour mobility and reduce regional unemployment disparity.

Chart A: Range of state wages growth
(Average weekly ordinary time earnings for full-time adults)

Chart A: Range of state wages growth (Average weekly ordinary time earnings for full-time adults)

Source: ABS cat. no. 6302.0 and Treasury calculations.