Talk to VCTA Teachers Day, Melbourne, Friday, August 28, 2015
Since it’s an important part of the course, I thought it might be useful if I give you an update on recent facts, figures and trends in the key indicators of what’s happening in the labour market, to help you keep up-to-date. I’ll be relying heavily on an article by professors Roger Wilkins and Mark Wooden, both of the Melbourne Institute, Two Decades of Change: The Australian Labour Market, 1993 - 2013, published in the December 2014 issue of the Australian Economic Review. We’ll start on the supply side, with participation, then move to the demand side and on to indicators showing the outcome of the interaction of demand and supply, such as unemployment.
Labour force participation
If you look at the overall participation rate since 1993, after the recovery from severe recession of the early 90s had begun, you see a reasonably strong rate of increase until the global financial crisis in late 2008. But this is misleading. All the increase came from rising female participation. Among men, the long-term downward trend continued, though at a slower rate of decline. However, in the five years before the GFC male participation rose a little. Since the GFC it has fallen by 2 pc points.
Note that some part of the fall in the male and overall part rates since the GFC is explained by demographic factors - that is, the retirement of the baby-boomer bulge. To get an indication of the size of this effect, Wilkins and Wooden calculated that if the age structure of the population - that is, the proportions in each age group - in 2013 was the same as it had been in 1993, the overall part rate would have been 2.2 percentage points higher. That is, population ageing seems to have reduced the part rate by a least 2 percentage points so far. We need to remember this when we look at the lower part rate in recent times - demographic as well as cyclical factors are at work.
Looking at the part rates for men, over the 20 years to 2013 the overall rate has fallen form 73.5 pc to 71.4 pc. The fall has been greatest for men aged 15 to 24, mainly because more are staying longer in the education system before starting work. Part rates for prime-age men - 25 to 54 years old - have fallen to lesser extents. But the part rate for men aged 55 to 59 has risen from its low in 1993 by 9 pc points to 80.8 pc in 2013. For those aged 60 to 64, it’s risen from a low in 1998 by 17 pc points to 62.5 pc. And for those aged 65 and above, the part rate has more than doubled over the 20 years to almost 17 pc.
The trend to early retirement - voluntary and involuntary - that started in the 70s and ran through to the early 90s, has really turned around since then and more men are waiting longer before retiring. A significant minority are continuing to work beyond age-pension age of 65, no doubt many of them part-time. I think this would be happening because many felt they hadn’t yet saved enough to live comfortably in retirement on an income above the pension, because more of them would be in better health and expecting to live longer than earlier generations were and more of them would have worked in less physically demanding jobs.
Turning to women, their overall part rate has risen continuously over the past 20 years, by almost 7 pc points to 58.6 pc. That’s true of all age ranges except for a very recent decline among those aged 20 to 24 - no doubt because more young women are going on to uni. By far the biggest pc point increases have been for those aged 55 to 59 (up more than 28 pc points to 65 pc) and 60 to 64 (up 30 pc points to 45 pc). Even the part rate for women 65 and over has risen by almost 6 pc points to 8 pc. The reasons for great proportions of older women remaining in the labour force longer include those I listed for men, plus the raising of women’s age-pension age from 60 to 65, and the fact that, with so many women now more highly educated and hence better paid, more of them, having returned to the workforce after having children, would want to stay there for the duration.
Looking to the future, there is plenty of scope for part rates among older people to continue rising. So, is the retirement of the baby-boomer bulge lowering the overall part rate or isn’t it? As you’ve probably realised, these days a fully-furnished economics teacher needs to know a bit about demography. Here we are balancing two separate effects. The fact that some older workers are choosing to retire later doesn’t alter the fact that population ageing is causing a lower proportion of the population to be of prime working age and a higher proportion to be of later age. This change in proportions guarantees that overall part rates will be lower. The decision of some older workers not to retire as early as they might have reduces the extent of the ageing-caused fall in the overall part rate, but is unlikely ever to eliminate it. This is because, although some choose to keep working, most do not. It’s unlikely we’d ever get to the stage where part rates among the over-60s were just as high as for prime-age workers.
Immigration
Roughly half the increased supply of labour comes from ‘natural increase’ (young people joining the labour force exceeding old people leaving it) and half from ‘net migration’. Immigration has been making a significant addition to the labour force since the end of World War II, and today accounts for more than a fifth of it. The main change since the turn of the century is that a smaller proportion comes on permanent visa and a much higher proportion on various classes of temporary visas.
In March 2014 there were 880,000 people on temporary visas with work rights, including 200,000 on temporary skilled (457) visas and 370,000 on student visas (not including many young people on working holiday visas). If all these people were actually participating in the labour force they would account for 7 pc of it. In addition there were 640,000 people on visas for New Zealanders. Since the early 2000s the annual intake of people on 457 visas has risen from 40,000 to 130,000 in 2012-13. These figures compare with an annual intake of skilled permanent migrants of about 130,000. In practice, many permanent visas are issued to people on temporary visa who decide to stay.
This shift to temporary migration has made the labour force more flexible in its ability to respond quickly to changes in demand for skilled workers without producing excessive rises in skilled wages. However, it has probably also reduced the incentive for employers to invest in training local workers. Note, too, that levels of net migration tend to vary with the strength of our business cycle.
Educational attainment
Now let’s move from the size of the supply of labour to its quality. The past two decades have seen continued growth in the proportion of the labour force with post-school qualifications. The proportion with a university degree has more than doubled from 12 pc to 28 pc. The proportion with trade qualifications has risen from 16 pc to more than 20 pc.
Employment
The best indicator of labour demand is the employment-to-population ratio - the ratio of total employment to the total population of working age (15 and over). Because the base is the total population rather than the labour force it will always be lower than the part rate. This employment ratio is getting more attention because it is now being included in the monthly labour force survey.
The employment ratio tends to vary directly with the business cycle. In mid-1993 following the recession of the early 90s it got to a low of 55.4 pc, but then slowly rose to a peak of 62.8 pc in late 2008 before the GFC, falling to 61.7 pc about a year later. It recovered temporally until late 2011 when it began falling again, reaching a low of 60.6 pc at the end of 2014. By July, however, it had recovered to 60.9 pc. The employment rate is reduced by population ageing, of course.
Unemployment and labour underutilisation
Unemployment represents the mismatch between labour supply and labour demand, the extent to which the labour market has failed to clear. The official unemployment rate is derived from the monthly labour force sample survey of more than 26,000 households. The official rate reached peaks of 10.4 pc in the recession of the early 80s and 11 pc in the recession of the early 90s. By 1993 it had fallen little, but over the following 15 years to early 2008, before the GFC, it got down to 4.1 pc. By mid-2009 it had risen to a peak of 5.8 pc, but recovered to a low of 4.9 pc in early 2011. It slowly worsened thereafter, but by mid-2015, however, it seemed to have stabilised at about 6 pc.
The official unemployment rate has not been tampered with by politicians, as many people believe, nor has the Bureau of Statistics changed its definition of unemployment for many decades. It is true that this definition is very narrow, in that anyone working as little as one hour a week is classed as employed. This narrow definition clearly understates the full extent of unemployment.
This explains why, some years ago, the bureau began publishing the under-employment rate, representing the proportion of the labour force that is employed but working fewer hours than desired. The underemployment rate did not slowly decline over the 90s and the noughties as the unemployment rate did, but stayed fairly steady over that period and has edged up a little during the 2010s. This means that, by the turn of the century, it went from being below the unemployment rate to being above it. Figures for the underemployment rate are published quarterly, for the middle month of the quarter. In May 2015 it stood at 8.4 pc, compared with the unemployment rate of 6.2 pc. Adding the two together gives a “labour force underutilisation rate” of 14.6 pc.
Note, however, that the underutilisation rate tends to go too far the other way by overstating the extent of the problem. This is because it counts the number of underemployed employees (and unemployed workers) without taking account of each person’s degree of underutilisation, whether they were a part-time worker wanting only an extra hour or two of work or, at the other extreme, were an unemployed worker seeking a full-time job. The bureau calculates a “volume-based” measure of underutilisation for August each year. In August 2014 it was 8 pc, compared with the “headcount-based” measure of 14.4 pc.
Wilkins and Wooden discuss another dimension of underutilisation: the extent that workers were not making use of all their skills and qualifications in their paid employment. Using data from the annual HILDA survey rather than the bureau’s surveys, another group of academics has estimated that, during the period from 2001 to 2006, 14 pc of workers were severely over-skilled (ie they made very little use of their skills and abilities in their current job), while a further 30 pc were moderately over-skilled. Wilkins and Wooden detected no trend in this measure over the past decade.
Youth unemployment
The rate of unemployment is always significantly higher among the young than among older workers. This partly because the young are more inclined to leave one job in the hope of trying another, because their lack of experience may make them less attractive to employers and because they suffer the brunt when the economy turns down. More recent employees may be the first to be laid off. And policies of reducing staff sizes by natural attrition usually involve suspending the annual entry-level intake at the expense of youngsters leaving the education system. The longer they remain unemployed, the harder it becomes for them to find a job.
In July 2015, the overall rate of unemployment was 6.3 pc, whereas the rate for 20 to 24-year-olds was 10 pc and for 15 to 19-year-olds was 19.3 pc. That is pretty much the highest it’s been since November 1997. Many of these teenagers would have quit school early, often with inadequate literacy and numeracy. Six years after overall unemployment reached its peak following the GFC, the overall rate rose by 0.3 pc points, whereas the rate for 20 to 24-year-olds rose by 1.4 points and for 15 to 19-year-olds rose by 2.5 points. This means one in five unemployed Australians is a teenager.
Note, however, that a teenage unemployment rate of almost 20 pc doesn’t mean that one teenager in five is unemployed. This is because it’s 20 pc of the teenage labour force (not the teenage population) and many teenagers are not in the labour force because they are in full-time education and not working or actively seeking work. It’s nearer one teenager in 10 that is unemployed.
Non-accelerating-inflation rate of unemployment
The NAIRU - or natural rate of unemployment - is the lowest rate to which unemployment can fall without shortages of labour leading to rising wage and price inflation. It’s regarded as the modern level of full capacity or full employment, or the lowest sustainable rate of unemployment.
The NAIRU changes over time with changes in labour market institutions and no one can say with certainty what its level is. Different economists do different calculations and reach a range of answers. However, the consensus among the econocrats is that the NAIRU is “about 5 pc”. But the only sure way to determine where it actually is, is to see how low it can get when the economy has been growing strongly before signs of wage pressure emerge. Considering the extent to which wages growth has slowed - to 2.3 pc a year - at a time when the official unemployment rate has risen only to a bit above 6 pc, my guess is the NAIRU may be nearer 4 pc than 5, maybe even less than 4.
The hardest question I have to answer from the public is how full employment could possibly be as sky-high as 5 pc. Wasn’t it less than 2 pc in the post-war years? Yes it was. But the structure of the economy has changed hugely since then, and most of the increase in the rate of full employment would be an increase in structural unemployment. Many people don’t realise that the unemployed tend to be unskilled and unsuited to fill whatever job vacancies are available. Figures taken from the HILDA survey by the Brotherhood of St Laurence show that in 2008, when the economy was booming just before the GFC and the unemployment rate got to a low of 4.1 pc, 45 pc of the unemployed were early school-leavers and a further 20 pc had gone only to year 12. That left 16 pc with trade qualifications and less than 20 pc had degrees. The problem is there aren’t as many jobs available for unskilled and low-skilled workers in the digital age as there were in the 1960s.
Overwork
Apart from the rise in part-time employment, the 1980s and 90s saw a marked increase in the proportion of men working very long hours. In the noughties, however, the proportion of male full-timers working 50 hours or more a week fell from 31 pc in 2003 to 27 pc in 2013. This has occurred despite an increase in the proportion of men working as managers and professionals. Among women, the proportion of hours worked changed little over the period.
Job insecurity and “precarious employment”
It’s widely believed that there is an ever-growing incidence of “precarious employment” - people in casual jobs, or on short-term contracts, or working for labour hire companies or temping agencies, or being cast adrift by their employer without benefits as supposedly self-employed. I know it’s widely believed among teachers that, these days, no one stays in the same job - even the same occupation - for long. More and more people are being made redundant. A young person leaving education can expect to have many different jobs before finally they retire at 70.
These perceptions may be widely held, but there is little evidence to support them. It is true we have a lot of part-time and casual employment in Australia - more than in most other rich countries - much of it done by mothers with young families and students who aren’t wanting a full-time job. And, these days, by people in semi-retirement. It’s also true that the number of part-time and casual jobs grew rapidly for several decades. It’s still growing, but much more slowly. According to Wilkins and Wooden, over the 10 years to 2013 the proportion of men working part-time has increased by 2 pc points to just under 18 pc, while the portion of women has been steady at almost 48 pc.
While most part-timers are also casuals, the two groups don’t overlap completely. The Bureau of Statistics defines casual employment as not receiving paid annual leave and sick leave. Its figures show that, for men, the proportion of casuals has been relatively steady since the late 1990s, fluctuating around 20 pc. Among women the share has fallen from about 31 pc to less than 27 pc.
The annual HILDA survey shows that more than two-thirds of workers were in permanent or ongoing employment in 2012, an increase of 1.5 pc points since 2001, when the survey began. HILDA suggests the share of labour-hire and temporary-employment agency jobs has fallen over that time from 3.7 pc to 2.7 pc. (It would be much higher than that in particular industries, of course.)
Nor can Wilkins and Wooden find any evidence that there’s been a shift away from employment to greater use of self-employment. Indeed, the bureau’s figures show the proportion of self-employed in the workforce has being steadily declining over the past 20 years, from 14 pc to 10 pc in 2013.
Then there’s the widespread perception that these days people are always losing their jobs and having to move on. When employers announce mass layoffs it invariably gets much attention from the media. When there’s nothing to announce it gets no attention. The bureau’s figures for average job duration and rates of job mobility show little sign that jobs have become less stable, according to Wilkins and Wooden. In February 2013, just 18 pc of the employed had been in their job for less than a year, down from 22 pc in 1994. In the latest figures, just over one worker in four had been in their job for at least 10 years, up from 23.6 pc in 1994.
How long people stay in the same jobs is determined by dismissals and quits. If jobs are becoming less secure you’d expect dismissals to be up and voluntary departures down. Both of these vary with the ups and downs of the business cycle but, even so, they’ve tended to decline. In February 2013, less than 3 pc of all the people who’d had a job in the previous 12 months had been retrenched. The proportion losing their jobs for any reason was 6 pc. About 10 pc of people had quit their jobs.
Earnings
The 20 years to 2013 have seen strong growth in real wages as the economy’s upswing continued. The real average weekly earnings of full-time employees grew by 38 pc for men and 37 pc for women. But the period saw no convergence between men and women in average weekly earnings, with the gap being 25 pc and 1993 and 26 pc in 2013. Other figures show the real hourly earnings of part-time workers increasing between 1995 and 2012 by 33 pc for men and 24 pc for women.
By contrast, between 1998 and 2013 the federal adult minimum wage rose by just 8 pc in real terms (compared with a 25 pc increase in real average weekly earnings over the period). Historically, the ratio of minimum wages to average wages has been a lot higher in Australia than in other developed countries, but this suggests it has been dropping back to the pack.
It also suggests that wage gains have not been uniform across the earnings distribution. Wilkins and Wooden use the bureau’s income survey to show that, comparing real earnings in 2012 with those in 1995, the percentage increase in the earnings of full-time employees is highest at the top of the distribution and lowest at the bottom. At the 10th percentile the increase was 15 pc, whereas at the 90th percentile it was almost 50 pc. So earnings inequality increased over the two decades. The Gini coefficient rose significantly from 0.26 to 0.29 for men (ie by 12pc) and from 0.21 to 0.25 (ie by 19 pc) for women. So, unlike in the United States, real wages at the bottom have grown to a fair extent, but higher wages have grown by a lot more, with middle wages growth somewhere in between.
The authors say the causes of this increased inequality in earnings aren’t well understood. In the US, increased international trade and skill-biased technological change are regarded as the most likely causes, with more emphasis on skill-biased change (ie computerisation that tends to reduce demand for less-skilled workers and increase it for high-skilled workers) than on trade (ie manufacturing and less-skilled service jobs shifting to developed countries). Both causes may have been at work in Australia. Ian Watson has argued that the increase in earnings inequality is the result of “neo-liberal” policies, such as deregulation, contracting out, reduced protection and privatisation. Another likely candidate is the decentralisation of wage-fixing with the move to enterprise bargaining.
Note that earnings inequality is not the same as income inequality. This is because, wages are not the only form of income, because the unit of analysis for earnings is the individual worker, whereas for income it’s usually the household (which may contain one or two earners, each on very different wage levels), and because measures of income inequality usually take account of the income redistribution brought about by income tax and cash transfers (including to non-workers, such as the aged, the unemployed and many sole parents). For these reasons, income inequality hasn’t increased to nearly the same extent that earnings inequality has.
Conclusion
The first of the big changes in the labour market we’ve seen over the past 10 or 20 years is the effect of the retirement of the baby-boomer bulge in significantly lowering the participation rate, which has been only partly offset by the reversal of the trend to early retirement, including more people continuing to work beyond the age-pension age of 65. Immigration continues to account for about half the annual growth in the labour market. The main changes have been the increased emphasis on skilled migration and on the issue of temporary rather than permanent visas, which has probably discouraged employers from doing as much training of local young people. Levels of educational attainment continue to rise. The narrowness of the standard definition of unemployment has led to publication of broader measures, which reveal a much greater underutilisation problem. The NAIRU is said to be about 5 pc, reflecting higher structural unemployment. The wide belief that jobs are getting more insecure and lasting fewer years is contradicted by the evidence. Earnings have become much more unequal, but other changes have limited the effect of this on income inequality.
