What’s happening in the labour market ?

With the Howard government focusing on industrial relations reform, it’s important to be well-informed about what’s going on in the labour market. Last time I posted on this topic, a number of commentators suggested that I had missed some important recent developments. In particular, regular commenter Derrida Derider supplied me with some useful stats that support this, so I need to begin with a revised assessment.

During the 1990s, I argued consistently, and I think correctly, that measured gains in productivity were being driven by increased intensity in the pace of work, longer working hours and a labour market that marginalised those unwilling or unable to put in long and intense hours, notably including older workers.

It now seems clear that most of these trends levelled out in the late 1990s, and went into reverse after 2000

Ross Gittins pointed this out in relation to working hours a couple of years ago and a graph suppled to me by DD shows the same for the employment-population ratio for men aged 55-64 (the trend rate for women was rising steadily, reflecting the first cohort of women for whom the expectation that married women should not work was not a relevant constraint).


And, as I never tire of pointing out, the period of above-average productivity gains ended around 2000. Since then, gains in multi-factor productivity have been at, or below, the long-term average.

Another area where I relied on casual observation and conventional wisdom is that of contracting. I accepted claims that an increasing proportion of workers were now employed as (often only nominally) independent contractors and that the proportion of wage earners was declining. DD points me to ABS 6291.0.55.001, Table 08, which shows no such trend for the period since 1984 (or any subeperiod) If anything there has been a slight increase in the proportion of employees, matched by a decline in employers and the disappearance of contributing family workers.

My take on all of this is not very original. Slumps shift the balance of power from workers to employers. Other things equal, long expansions shift power the other way. Even in the US, where ordinary workers have been losing ground for decades, the last few years of the boom of the 1990s saw increases in real wages across the board.

But “other things equal” is important. Now that the Howard government can do what it likes in IR, employers will be given the whip hand, and many will be keen to use it, particularly if the housing and commodities booms subside. More on all this from Tim Dunlop, Daily Flute, Alex White and WSACaucus

50 thoughts on “What’s happening in the labour market ?

  1. John Humphreys makes an important point – productivity affects participation. But we can’t say a priori whether the income effect dominates the substitution effect. In the 70s people were worried what everyone would do with their leisure – they expected people to become so productive they wouldn’t bother to work more than a few hours – while John seems to think that the prospect of more earnings will make people want to work longer. The history of working hours and participation since the 70s suggests he might be right (more of the working age population is now working than ever before in our history).

    But participation affects productivity too – the more low-productivity workers have jobs, the lower average labour productivity will be (that’s why French labour productivity is higher than in the US – the French keep their marginal workers out of a job).

    James, amongst groups that have a lot of people with potential marginal attachment to the labour force (ie they have choices such as early retirement) the EP ratio is a much more reliable indicator of what’s happening than the PR. And EP ratios amongst 55+ men (and women) have risen 11 percentage points since 1993 – 8 ppts points of which is a reduction in measured unemployment (the residual 3 ppt is the recovery in the PR).

  2. John Quiggin’s reasoning appears to be that workers were worked harder and longer in the 1990s, hence there were measured productivity gains. I don’t know how to measure work intensity, and will leave that up to the experts. Nevertheless, average working hours actually declined between 1992 and 2002:


    Average weekly hours worked by employed persons fell from 34.7 to 33.7 hrs/wk. Average “full-time hours” worked was constant at 40.8 hours per week, while average “part-time hours” worked rose from 15 to 15.9 hours. Average weekly earnings for full-time adults rose from $621/wk to $931.50/wk.

    Perhaps work intensity (however measured) explains higher labour productivity, because I can’t see the evidence for longer hours.

    Another interesting change over time has been the fall in real-unit labour costs. I’d like to hear John’s theory for the decline in compensation to labour relative to other factors of production. There is a discussion on this theme over here:


  3. DD

    I didn’t notice you’d replied to my post chez Baboon. Thanks.

    As you say, the line between unemployment and non-participation is both blurred and volatile for older workers, who can’t decide whether they’re actively seeking work or not. This is important because it means the unemployment rate itself is not a good guide to either (a) labour market tightness and the likelihood of wage inflation, or (b) how bad things are for these workers.

    But how is this relevant to the matter at hand?

    John said a declining participation rate was evidence that the labour market no longer meets the needs of older male workers. Since then, we’ve all agreed that it hasn’t declined since the late eighties, but I maintain that at best it hasn’t improved after declining sharply in the seventies and early eighties. But you go further and claim that the labour market is in fact serving older workers pretty well, on the basis of a growing employment rate, which you say gives a more meaningful picture.

    Of course, if the employment rate is rising faster than the participation rate, this implies that unemployment is falling. And this is good, but it’s a cyclical effect, and none of us is talking about that. When we adjust for the cyclical effects, say by comparing 2005 with 1990, and we still find that employment has increased more than participation, what does it mean? Presumably that there has been some fall in structural unemployment.

    As I read all these figures for the male 55+ group, taken together they say that, relative to the late 1980s, part-time employment has risen a little and unemployment has fallen a little, with full-time employment and inactivity about constant. So some older people who might have been unemployed even in the 1980s boom are now in part time work. I guess that’s good, but it’s not great, and we still have the problem with the younger cohorts.

    Finally, I think John Humphreys’ claim is both implausible and untestable.

  4. John, I will just make a quick comment about ABS stats and contracting. The data set you mention, ABS 6291.0.55.001, Table 08 does not contain enough information to make determinations as to the prevalence of “contracting.”

    I will make two assumptions here. First, you refer to the fact that so-called independent contracting arrangements are often not genuine, so I presume you are using the term contracting to refer to non-standard employment arrangements? In any case, that is the usual use of the term in this sort of discussion.

    Second, I presume that in your interpretation of ABS 6291.0 you used the category of own-account worker as a proxy for contractor?

    The problem here is that own-account worker is not in fact a valid proxy for either true independent contractors or for non-standard employees in general. There are two reasons for this. One is that workers in these arrangements often don’t understand who their employer actually is.

    The second is that it is now common practice for labour hire placements to be structured as one dummy company providing the workers to the labour hire firm for on-placement to the true employer. In these arrangements, the dummy company is the nominal employer of the worker and, for all intents and purposes, the worker is recorded as an employee of that dummy company. Labour hire practice adopted this tactic as a defence against court cases that found labour hire firms or the end employer were the actual employers of workers.

    Contracting as discussed here is intimately tied in with casualisation, and the definitive work on this topic in Australia is Watson and Buchanan’s Fragmented Futures. A more appropriate way to detect non-standard employment arrangements is to consider whether leave is included. Under that criteria, Buchanan and others find that casual/contract employment increased from 16 percent in 1985 to 27 percent in 2002. (Watson p69)

    This is generally supported by other data. For example, recruiters enjoyed a 31 percent increase in revenue from 1999 to 2002, and a 37 percent increase in placement numbers. (ABS 8558.0)

  5. While I’m here, I’ll mention what I think the real impact of the IR reforms will be.

    First, by raising the exclusion criteria to 100 employees, the reforms effectively apply to all corporations in Australia, even the huge ones. This is because 100 is large enough to conveniently siphon off workers onto placement contracts. That is, the banks, Qantas and anyone else can start shifting non-core workers into numerous dummy companies that then provide labour to the original employer, as per the increasingly prevalent labour hire arrangements.

    Second, forthcoming FTA’s with China and certainly India will almost certainly contain GATS Mode 4 type provisions for movement of natural persons. The effect of those provisions will be to saturate Australian employment markets and drive down wages and other conditions. We are in fact taking a step towards third world inequalities.

  6. SH’s comment at 7 is inaccurate. While working longer for less is indeed an increase of productivity, you can also get productivity increases when both go up – or when both go down. It’s just that the ratio needs to change.

    Firm’s productivity affects their incentives, and other costs of drops in short term production get externalised. As well as which, as, when and if things do trickle down into more employment, it’s not the same people who are the gainers (even the Victorians saw this – read Disraeli’s “Sybil”). Also, further waves of “improvement” can make a sort of shock wave effect, overwhelming trickle down.

    All of which leads to some of the material I touched on my publications page, not just the Kim Swales stuff, so I’ll leave it there due to lack of public interest.

  7. John:
    I don’t understand why you are critical of my comments.
    Unless I am mistaken, you are saying that productivity increased during the period in question because workers were made to work harder: work intensity, is the term used.
    There is really nothing new in this theory as it goes back (if I recall my economic history correctly) about 200 years ago when there was vigorus debate on the marginal utility of labor. Those who believed (I think started by Ricardo) that profit was derived through the exploitation labor were widely discredited. This argument proved to be fallacious. Unfortunately we continue to see this argument rise from the ground like a dead body. Labor cannot improve productivity without capital. If this wasn’t the case money would be pouring into places like Tanzania where labor is plentiful and I am sure if an employer used a whip it would easily be overlooked.

    Howard’s labor deregulation.

    If you want to observe what regulated labor market looks like let’s debate what unemployment rate looks like in France and Germany. During 2004 world growth was 5.5%: probably one of the highest rates seen for decades. German unemployment actually rose during this period. France didn’t quite (but almost) achieve this stupendous result. Unemployment there just hung around 10.5%. It would be wrong to argue that the appreciation of the Euro had much to do with Germany’s plight because unlike France or Italy, Germany’s trade is mostly inside the EU. The ONE and ONLY reason why unemployment is so high is because of the constipated labor market in these countries. If employers can’t fire they won’t hire. Daimler Benz and BMW have not built a new plant in Germany for 15 years and I bet they won’t, until labor laws are changed.

    (I am bringing this up because I saw a few posts criticizing Howard for making these proposed changes).

    It is insane to decry these proposals if you really are on the side of the workingman. What we want are not laws governing hiring. We need labor markets, which price labor to its highest rate possible and the only way we can achieve this is by having plentiful jobs. Those wanting restrictive labor laws to remain are really demonstrating selfishness cloaked up as altruism.

  8. Give it up S brid,

    In your first point you are implictly equating productivity with profitabilty, Wrong, unless you are representing a class instead of a technical point.

    And your second point is crap. First, unemployment in the US is a serious issue. Yes, another jobless recovery. And if you’re going to quote Germany, how can you possibly exclude communist China? 10 per cent average annual growth, and stick your Washington Consensus up your arse, thank you very much and how’s your father.

  9. CS,

    There’s very little communism at work in China’s GDP growth rate – quite the reverse. And unemployment is much worse in Germany than it is in the USA.

  10. S. Brid

    Either you didn’t bother to read the paper on productivity that John linked to, or you are out of your depth. In the unlikely event that you are capable of grasping the point, let me put it in twenty-five words: when they talk about productivity, unless otherwise specified, economists don’t mean output per man-hour, but rather multi-factor productivity, which is adjusted for capital intensity too.

  11. CS
    Hope I my post didn’t offend you in any way as you seemed quite angry. I will try to help you understand where you are wrong as best I can.

    1. Wasn’t equating productivity to profits in any way? I was talking about why I thought John’s comments on work “intensity� were wrong. Please look at the earlier post to see my comments on productivity.
    2. I really don’t understand what you are saying when you are talking about “classâ€? etc. Did you forget to add anything to that, it’s incomprehensible?
    3. I agree with you about the point of unemployment being a serious issue in the US. At 5.2% it is serious. The only comments I would make to that is that the US has created over 20 million new job in the last 20 years. Germany, nil. The US has further to go to get unemployment down, only Germany is over double that now and direction still points north. I would also make a comment about the resilience of the US economy. Imagine taking a hit like 911, going to war, the tech crash and the place continues to grow at a good clip. Please, I am not making any comments about US foreign policy etc. I am just making an observation on its economic performance. That’s all.
    4. Yes of course China is growing at 10%. Funny how liberalizing capitalist based economic reforms have that effect isn’t it. I am glad you brought that up. The parts of the Chinese economy, which are growing massively, are those where there is almost no interference or very restrictive labor laws.
    5. I don’t know why you brought up the “Washington “consensus� charge. Would you kind explaining this as it looks like you left something out.
    6. I would have thought that referring to my rectum and father in the same sentence would qualify as rough language.


  12. By the way, Steve Edwards, no one is claiming that longer hours increases productivity. Unless they are unrecorded hours (and unpaid?), which I think is a possibility worth investigating.

    And it’s nice to hear from Tony again.

  13. Hi James:

    unless otherwise specified, economists don’t mean output per man-hour, but rather multi-factor productivity, which is adjusted for capital intensity too.

    Just did a quick google for the exact defintion of productivity. Typed in “defintion of productivity”. see link http://www.google.com.au/search?hl=en&lr=&oi=defmore&q=define:Productivity

    All expose the definition as being the “amount of output per unit of Labor”.
    I described it as “the number of widgets produced in set period of time”.
    More or less the same thing.
    Also checked a dusty Samuelson. Pretty much the same thing.
    Please let me know if this is wrong as I would like to correct myself if that is the case.


  14. But James, John Quiggin appears to be saying just that:

    “During the 1990s, I argued consistently, and I think correctly, that measured gains in productivity were being driven by increased intensity in the pace of work, longer working hours and a labour market that marginalised those unwilling or unable to put in long and intense hours, notably including older workers. ”

  15. So he did, Steve. I should have checked. Presumably he means the extra hours are unofficial and unrecorded, as I suggested. Alternatively, he meant that the extra hours are more productive at the margin, and of course this would be a problem only if the employees were bullied into putting in the extra hours (which I suspect is often the case, but that would need be proved). Let’s see if a clarification is forthcoming.

  16. I made the point about unrecorded hours, including such things as the loss of breaks, in a number of papers during the 1990s.

  17. James: Finally, I think John Humphreys’ claim is both implausible and untestable.

    I didn’t make any claim. I think James just enjoys dissagreeing with me — even when I’m not saying anything. 🙂

  18. No, Steve Edwards, if you work an extra hour your production has risen. your productivity (ie output *per hour* worked) has probably fallen a little (production is generally subject to diminishing marginal returns).

    There is genuine confusion here about labour productivity versus multi-factor productivity. Lots of economists (esp labour economists) do use the term “productivity” as shorthand for “labour productivity” – the partial differential. Others (esp macroeconomists) use it as shorthand for MFP – the total differential. JQ in his scribblings is in the latter camp. And S Brid has not yet grasped the difference, probably because of this inconsistent shorthand usage.

  19. DD, precisely because multifactor productivity is a bundle, it is a vector, not a total. And when you think, well, that was just verbal shorthand, isn’t that precisely what the others were using? It’s not good enough for precision work, of course, but it is adequate for this sort of general illustration.

Comments are closed.