Bleg: Can you explain this graph? (changes in male full-time employment)

As the graph below shows, the proportion of men in full-time work has fallen over time. Every recession the proportion falls sharply and in each recovery it fails to bounce back to its pre-recession level.

When I show people this graph they often offer explanations — it’s population aging, kids staying in education longer, or married men staying at home to look after their kids. But when I look at the data more closely, I’m still puzzled.

Occasionally people will interpret my question is a plea to return Australia to the 1950s where men won the bread and women were detained in fibro boxes surrounded by neatly trimmed lawns. I’ll be reminded that men are not a disadvantaged group and that women’s wages still lag behind men’s.

But despite the helpful explanations and the risk of having my motives misinterpreted, I’m still curious about the data. I suspect it might have something to do with structural change in the labour market — shifts in the demand for different kinds of skill. And I imagine these changes have affected some groups of women too. Perhaps it’s got something to do with the decline of industries like manufacturing?

In the end, I’m not entirely sure how to explain this. Can Troppo readers help?

More graphs below the fold.


Sources: 6202.0 – Labour Force, Australia, Apr 2011 & 6291.0.55.001 – Labour Force, Australia, Detailed – Electronic Delivery, Mar 2011.

Update: An obvious question is: If men aren’t working full-time, then what are they doing? Here’s a chart showing changes in the labour force status of men aged 25 to 34.



Update 2:
Here’s a chart for Tel (see comment 4). It shows how the percentage of men aged 25 to 34 not in full-time employment has changed over time. It’s broken down by labour force status — NILF, part-time employed and unemployed.

Update 3: A related issue is the increase in the proportion of men working in casual positions. Earlier this month the SMH’s Paul Bibby reported:

Men have increased as a proportion of the casual workforce from 39 per cent in 1992 to 44 per cent in 2009. There is also a growing prevalence of casual work among the so-called prime working age categories, those between 24 and 54.

It suggests the start of a fundamental shift away from the classic notion of ”male work” as a permanent salaried position based on a 40-hour week.

”You’ve now got quite a large group of prime working age men on casual arrangements,” Dr Campbell said.

There appears to be a growing divide within the casual workforce between the haves and have-nots. More than a quarter of the casual workforce – about 605,000 casuals – are working full-time hours.

At the other end of the spectrum is the large and growing number of underemployed casuals.

Update 4: Conrad (comment 2) links to a recent article in the Economist: Why ever fewer low-skilled American men have jobs. According to the article:

The main reason why fewer men are working is that sweeping structural changes in rich economies have reduced the demand for all less-skilled workers. Manufacturing has declined as a share of GDP, and productivity growth has enabled factories to produce more with fewer people. Technological advances require higher skills. For the low-skilled, low demand has meant lower wages, both relative and absolute. This in turn reduces the incentive to find a job, especially if disability payments or a working spouse provide an income.

… Broadly speaking, this is a common story across the rich world: in virtually all OECD countries male employment rates are lower than they were 40 years ago, and the decline in America’s rate since the 1970s is similar to others in the G7. But in America the timing has been different: the fraction of men in work has fallen especially quickly in recent years.

Update 5: In comment 16 desipis raises the issue of underemployment. Here’s a graph from 6265.0 – Underemployed Workers, Australia, Sep 2010.

This entry was posted in Blegs, Economics and public policy. Bookmark the permalink.
Subscribe
Notify of
guest

36 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
conrad
conrad
12 years ago

Some nice cross-cultural data and a discussion of it on a very similar issue was presented in The Economist quite recently. See this link. Interestingly, there is quite some variation between different countries.

Rafe
12 years ago

Nice to see that men are moving from fulltime to part-time work so they can do more housework and childminding.

Tel
Tel
12 years ago

Could you please re-stack the bottom graph (top to bottom)?

* Not in labour force
* Part-time employed
* Unemployed
* Full-timee employed

I think that would neaten it up a bit.

Harlequin
Harlequin
12 years ago

Death of manufacturing? As a consequence of tariff reduction. Exacerbated by the recession in the early 90s. Just guessing.

Richard Tsukamasa Green

I’ve been interested in graphs like this since I first saw one in Honours Labour Economics. The fact that it is so stepped is what makes it so intriguing. It doesn’t look like a long term trend with large breaks at recessions, it looks like a flattish trend with large breaks that are neither extensions or reversals of what was happening leading up or following them.

My lecturers were of the belief that the recessions caused a structural increase. When men stopped working they gained both a stigma and lose skills. With gaps in their resume they would get a lemon tag, future employers would figure there must be something wrong, otherwise someone else would have hired them. This is more true for men than women because being a part time worker or stay at home partner is not as socially accepted for men and staying with kids would be a “bad excuse” in the eyes of employers. On the other hand recession is a pretty good excuse. Also unemployment would degrade both explicit skills (as well as being unable to upgrade on the job skills as the workplace changed) and tacit skills (punctuality etc.). I am unsure why this would disproportionately affect men (unless unskilled work is “feminine”).

The problem I see though is that we should see a slow, and probably incomplete, reversal between recessions as new men come into the workforce and the stigmatised/skills eroded age out of the working age population. Additionally, as unemployment got lower employers would have slightly less capacity to sort by stigma and take what they could get (I stress slightly, a 5% NAIRU is still alot of people). The age trend would be mitigated by the boomer demographic bulge, since anyone affected would only now be aging out of the relevant population.

But we only see any kind of reversal trend after the post GST slump.

Another explanation could be that long term structural trends are just making large parts of the male workforce unemployable because of changing skills as manufacturing declines and technology changes and also the fact that employers increasingly have the option (and inclination) to employ more capable candidates (women) that had not been in the workforce or had been excluded. This could explain the stepping effect since good times could obscure the structural changes and recessions make them clear so that the newly unemployable would lose their jobs in cohorts.
But this would also run into problems because I’d still expect to see a long run trend between recessions. Men should have been losing/trying to change their jobs all the time and then prove unable to find another. Attrition would lead to a trend with accelerations at recessions. I can’t see that though (and I don’t buy that workers can so neatly go from employable to completely unemployable ).

And now I’m intrigued by that last graph since it does show that the steps are in full time employment, and that unemployment declines post recession, but the decline is shared by movement both into non participation and part time work. This might support the first hypothesis if some of the stigmatised/skills eroded are just giving up and others are eventually accepted into part time jobs which are less skilled and have less stigma.

But honestly I really don’t know.

Harlequin
Harlequin
12 years ago

There was also increased participation of women in the workforce. Or something. Honestly, others have done this subject justice. You could try George Megalogenis’ “The Longest Decade”. He covers off this subject and many others.

And anyway, it’s a good read.

conrad
conrad
12 years ago

“Or maybe I’ve found the right book but I’m just too dim to understand how it answers my question.”

Or, like many problems such as this, perhaps the real explanation relies on 7 different variables, 4 interactions between them (including a 3-way interaction), with 3 mediated and two moderated links (or whatever complex model you can dream of). This means that once the model becomes complex enough to explain a reasonable proportion of the data, almost no-one will understand it (at least in a single blog post!).

For example, a case like this would have to be somehow summarized in your graphs, and would boil down to someone sitting in one category: “In the last recession, I decided to leave my job as I wasn’t getting pay rises anymore. However, I started finding it difficult to find a new job, but this wasn’t such a problem because I found out that I actually enjoyed spending time at home and I could live happily off my wife’s salary, and she didn’t mind at all. After 3 years of this, I decided that I did in fact want to work part-time, but found out that this wasn’t viable due to EMTRs. Despite this, since I was offered an enjoyable job, I ended up working half the amount of hours I wanted, but found this just about right”.

Or perhaps there really is a simple explanation.

trackback

[…] 2: Here’s a chart for Tel (see comment 4). It shows how the percentage of men aged 25 to 34 who are not in full-time employment has changed […]

Jim Belshaw
12 years ago

Don, I can’t give you a rigorous answer, Just a few observations based on experience & reading. Richard@7 picked up on some of the issues.

Don Aitkin’s The Way we Were provides one answer; this was a study of social change in Australia through the eyes of the Armidale High School LC class of 1953. A good proportion of that class in permanent employment with super took early retirement because they became disillusioned with the changes taking place in the workplace. Given gender structuring, that took a first slab of men out of the workforce.

Don also commented that the people who lost jobs were not the same as those who gained jobs. That fits with my experience.

In the mid nineties I did some work as an outplacement consultant. Most of those I saw were men, generally older, who lost their jobs as a consequence of process reingeering and restructuring. Most had worked for extended periods, often all their life, for one company. Those aged fifty plus struggled to get new jobs. Some dropped out, others ended up with part time or contract work.

Patterns varied greatly across the country and between groups. In the country, farm work diminshed greatly from the early 1950s, then middle class jobs were cleaned out during the economic changes of the 1980s and 1990s. The new jobs that came were of a lower level and fell in occupational groups traditionally dominated by women.

In NSW, Aboriginal men were especially badly affected because they were less well educated, lived in regional areas and worked in those jobs that were most prone to change. One Aboriginal work program moved Aboriginal people to Newcastle to work in manufacturing where there were unskilled and semi-skilled labor shortages. This was quite succesful, but quickly went into reverse as economic restructuring trimmed manufacturing jobs. Aboriginal families moved back home; if you are going to be unemployed, then it may as well be in area where costs are lower and you feel at home.

Generational unemployment among NSW Aboriginal people emerged in a way that hadn’t been seen for decades.

There were other changes as well, including the growing dominance of women in the educational stream.

This has become a long comment. I guess my point is that the changes that Don Arthur pointed to are multi-factoral. The stats are a summary of a whole series of smaller changes that affected different groups in different ways and at different times. You can’t tell this from the stats. You have to drill down, using qualitative evidence.

Jim Belshaw
12 years ago

Thanks for finding the earlier post, Don. I don’t know enough about HILDA to be sure. It seems to have a price that I can’t afford!

desipis
12 years ago

It suggests the start of a fundamental shift away from the classic notion of ”male work” as a permanent salaried position based on a 40-hour week.

At the other end of the spectrum is the large and growing number of underemployed casuals.

I was poking around in the ABS data, and by plotting the underemployment rates of males I saw something that mirrors the dramatic steps seen in the first graph. While the steps are not of the same magnitude, the lack of any significant decay indicates that a fair portion of workers affected were and remain dissatisfied with the shift. I suspect the steps reflect employers taking advantage of some combination of workplace regulation changes and a vulnerable workforce in order to push away from regular full-time positions and entrench a culture of casual & part time positions.

Looking at the total numbers, it seems that the unemployed not only have to compete amongst themselves for any new jobs but against a greater number of people active and proven in the workforce who are seeking more work.

Jim Belshaw
12 years ago

Don, those graphs are quite fascinating. In focusing on younger men 25-34 you show broad structural trends quite clearly – the rise in the proportion not in the workforce plus the rise in part time employment. The last graph, underemployed workers as a proportion of part time workers, is harder to interpret.

In regard to despis’s comment @16, if you plot the proportion of casual and part time positions over time you can see the steady increase. Now you need to add in contract here as well.

One of the issues with contract work, and I speak from experience since I have chosen to do contract work over the last year or so to free time for writing, is the downtime between jobs. In an imperfect market, and the job market is imperfect, search takes time.

I am not sure to what degree contact affects male-female divides, although I think that more men are involved in contract work.

Richard Tsukamasa Green

Don – I should have mentioned that I was sure you knew all of this. I started typing it out to restate what I knew to myself (just in case I skipped something by using a jargon shorthand like hysteresis) and then figured I may as well leave it for anyone else coming along despite the fact I had come up with nothing new.

Andrew Norton
Andrew Norton
12 years ago

A figure by education level would be useful. Unfortunately unlike for other trends the ABS does not make it easy for you.

Tel
Tel
12 years ago

Thanks for the “Update 2”.

Note that the 1982 recession and the 1992 recession “step” features do not reflect through to interfere with the steady and smooth rise of NITLF and part timers. This suggests that short-term volatility in the market is primarily soaked up by the unemployed, although perhaps the part-timers get more or less shifts as economic activity changes pace thus soaking up the volatility in a way that gets hidden in the statistics.

Interestingly, the more recent 2008 step is different, because it does echo down though all the lines (even though the total magnitude of this step was smaller) suggesting a different mechanism at work. This is also the sharpest of all the steps.

Julia
Julia
12 years ago

Here’s a stab at a social rather than an economic explanation for one cohort. This only holds for those younger men in the 25-34 age group, and applies to those for whom changing work patterns are a choice and not forced onto them. It’s not based on data but on my own observation.

I know young men who have deliberately chosen not to take full time work even when it was offered to them at quite attractive rates and in conducive social settings. This is because they value discretionary time use over security of employment, or that’s what those few I have informally spoken to say.

My own theory is that these young men grew up in a time of prosperity and so have no particular use for security. They are used to feeling a bit exploited by work, and have no great affiliation for any workplace. They are I think, reacting to the eighties severance of company loyalty to employees in favour of efficiency. Therefore they are not loyal back, since trust and loyalty are two way streets, and this makes discretionary allocation of time even easier. So it would be interesting to know what this group is doing with alternative uses of time, leisure, travel, study, being annoyed at Sony for its gaming failure etc.

Dehne Taylor
Dehne Taylor
12 years ago

Bob Gregory has been interested in this for at least the number of years on your graph. Suggest you read some of his stuff or send him an e-mail.

derrida derider
derrida derider
12 years ago

I’ve done a lot of work over the years on male employment in Australia. 6291.0.55.001, 6291.0.55.003, 6224.0.55.001 and 6310.0 are now old friends and all are great fun to play with if you’re handy with a spreadsheet – well done the ABS. For those interested more in sociology rather than economics, I particularly recommend 6224.0.55.001.

Pretty well all the things driving male employment have been covered above. The simplest thing, though, is to look at that first chart and divide it into pre- and post- 1993. Pre 1993 it was all about recessions and industry restructuring. Post 1993 it’s mostly about simple population aging, though rising education retention (and hence increasing rates of part time work***) has mattered for the younger age groups.

Employment rates among older workers – both male and female – are really driven by cohort effects. As a broad generalisation the baby boomers now in their 50s are far more skilled and far healthier than their parents were at the same age. Hence they’re both more employable and more interested in staying on (because they can earn decent wages). In 1978 70% of men aged 55-64 were working. In 1992 just 53% were. Even in 2002 only 60% were. But in 2011, 70% are again and that number shows every sign of continuing to rise.

Phased retirement seems to be becoming popular though; most of that increase in part time work at the expense of full time work is a choice – a choice not available to previous cohorts. Also, there’s actually something in Rafe’s comment @3.

*** Australia has the highest proportion of youth working part time while studying full time in the world. It’s a product of the extreme stinginess of our student allowances, especially the still-ferocious parental means test. In turn, that’s part of the price we pay for our quite misplaced obsession with “middle class welfare”.

derrida derider
derrida derider
12 years ago

Don @15, much as I am inordinately proud of HILDA (I wrote the initial paper in the bureaucracy proposing it, I wrote the CabSub to get it funded, and I was its first Project Officer) it is not great for this particular job because:
a) it aint run for long enough (yet).
b) its sample size means you can’t find enough of small sub groups (eg baby boomer men who voluntarily retired because of disillusion with work) to do rigorous statistical work on them.

On education and work in differing cohorts, the best you can get are 6227.0 and (to a lesser extent) 6278.0. The publications here, though, don’t tell you all that much. They don’t do much in the way of time series and also don’t break some of the most vital variables up by age. You need to use the CURFs – which is expensive, extremely time consuming and needs proper statistical programming expertise.

conrad
conrad
12 years ago

“But what’s going on with younger men — men 25 to 34? Why is the fall in full-time employment for this group so persistent? Why didn’t it bounce back to the late 1980s level after the last big recession?”

I think a big factor is likely to be that most job growth has been for people with degrees (despite the rhetoric from many people and the government — obviously they don’t care about ABS figures), and females have been outcompeting males at university for ages. The link below has the older figues that should relate to that generatation where you can see this change, and the newer figures look similar — basically the difference between males and females has grown more slowly since then — it is very slightly under 60% now if you don’t include OS students.

http://www.deewr.gov.au/HigherEducation/Publications/HEStatistics/Publications/Pages/TimeSeriesStudents.aspx

So my bet is that many are not working full-time simply because they can’t get a job or because they are not willing to take a full-time job that does not appeals to them. However, because part-time work is much more available than two decades ago and since they need some money just to survice, I assume many take this alterantive either because they can (Julia’s suggestion) or because they are essentially forced to (since a lot of the growth in full-time jobs has been for people that need a degree).

Julia
Julia
12 years ago

Following on from Conrad, my tiny observational database also suggests a further extension of that thinking. Of those males in the 25-34 age group who can choose, the idea of putting together a “portfolio” approach to work seems to have appeal. I don’t know whether the statistics show if one person has more than one part time or casual job, or a mixture, but I’m betting the figures on multiple part time/casual jobs per person, if they exist, are going up, especially in that age group of males. It may also be the case for females as well but hidden under general female part time work.
I also suspect that as Jim Belshaw indicated @18 that the “lumpiness’ of work is going up as opportunity frequently clumps together, and does not line up precisely with availability, and search takes time.
A spin off from this is short planning horizons. Full time work means you can predict what is happening in several weeks or a year. This more opportunistic approach means you cannot. However the response to that is not to try harder to plan but to increase opportunism. Anyone with experience of that Gen Y group knows they accept an invitation till something better comes along socially as well as work wise.

(PS derrida derider, do you have a source for your observation @26 “*** Australia has the highest proportion of youth working part time while studying full time in the world “? Its an interesting one. )

Chris Lloyd
Chris Lloyd
12 years ago

“I’ll be reminded that men are not a disadvantaged group.”

You should remind them even more forcefully that men are indeed a disavantaged group. They lag behind women on every known measure of physical, and psychological well-being. Life expectancy (males have a higher rate of death at every single age, including year 1); suicide rates; criminal incarceration; victims of crime; depression; drug and alcohol dependence. You name it. The only metric they win on is money in the bank. Which suggest to me at least that “happiness research” though hard to do, is well worth doing.

James Rice
12 years ago

Although men do have greater access to positions of great power at the upper echelons of society (in government, business, etc). Nevertheless the aspects of quality of life you mention (life expectancy, suicide rates, crime victimisation, etc) are obviously important – and more could be added to the list (education, access to a range of supportive social relationships, homelessness, etc). Inequalities in these aspects of quality of life should be taken seriously, as should inequalities in wage rates, household work, and so on.

As an aside, money in the bank is also complicated, both because of sharing of income within households and because there is some evidence that, while men tend to receive more income than women, women tend to have more control over this money than men. In an article in the most recent issue of the Journal of Sociology, Supriya Singh and Clive Morley analyse data on married men and women from the 2006-2007 HILDA survey to look at gender and the control of money within marriages. They find that married women are more likely than married men to have their own separate bank account (married women are more likely than married men to have both joint and separate bank accounts, while married men are more likely than married women to have joint bank accounts only). Singh and Morley emphasise that they believe that this gender inequality in access to a separate bank account is only marginally relevant to the control of money within marriages (based on a qualitative study of a convenience sample of 108 people drawn from Singh and Morley’s own personal and professional networks, in which women were overrepresented). Despite this, Singh and Morley’s analysis of HILDA also suggests that within marriages decisions about large purchases are more likely to be made by women than by men (according to the reports of married women, in 46 per cent of marriages decisions about large purchases are made by the woman, while in 16 per cent of marriages these decisions are made by the man). Moreover, Singh and Morley’s analysis also suggests that within marriages day-to-day spending and paying bills are more likely to be managed by women than by men (according to the reports of married men, in 34 per cent of marriages day-to-day spending and paying bills are usually or always managed by the woman, while in 24 per cent of marriages these activities are usually or always managed by the man). (These latter results can be derived from results presented in the article, but they have to be derived. For one reason or another, Singh and Morley themselves do not discuss these results.)

Paul Frijters
Paul Frijters
12 years ago

Don,

nice graphs. This is obviously a core topic amongst academic labour economists. You can find very similar graphs for New Zealand and most OECD countries.

My best guess? Apart from the demand shifts that made more men fundamentally unemployable (and that have little to do with manufacturing), I would point out that in the 60s and 70s there was a lot of implicit welfare inside organisations. Dead wood was protected as communities looked after their own. It is that local welfare arrangement that has changed and been replaced by centralised systems that I think is primarily responsible for the long-term decline in the employment of men. The shift to part-time is more cultural and is probably connected to students and old people who either have something else to do or find it hard to work whole weeks.

derrida derider
derrida derider
12 years ago

Paul makes an interesting point about how declining productivity with age used not to lead to unemployment amongst older men because of organisational cultures. It squares with some anecdotal evidence I have. I wonder how you could test this properly, though – I doubt you could get the hard data needed on, say, employment practices in IBM in 1969 in order to compare it with today’s.

Julia – on the students working part time, it came from an OECD publication in about 2000. I’ll try and dig it up.

Cameron Murray
12 years ago

I know I’m late to the conversation but I thought that the answer was obviously a combination of many factors

1. Longer periods in education (that creep into the >25 years age bracket to explain Update 1)
2. The nomadic yearn for travel of the young cohort – which fits nicely with more part time work and more males out of the workforce.
3. Stay at home Dads – if women are working, men are more likely to spend some time as the primary carer and out of the workforce.
4. Generous welfare that means those on the fringes of disability or ability no longer feel the need to work
5 Demographics (to explain much of the first graph) – ie a smaller proportion of men at working age.
6. More career changes, hence more study breaks over the career.
7. Data accuracy – perhaps there is a slowly growing informal economy in wich welfare recipients are partaking?

But I must say, there is a lot of fuss over what appears to be a very small change for wokring age men – for the 25-34 year old cohort a drop from 95% of men in the workforce to a ‘mere’ 92%! Clearly much the explanation for the whole of population trend rests outside this range – at transition to retirement and the transition from study to work.