Government report cheers for wage cuts

There’s been a lot of discussion recently about stagnation in real wages and the decline of the labour share of national income. In a recent Senate Submission, I made the point that there is nothing surprising about this

For the last 40 years, changes in labour market regulation have been almost uniformly anti-union and anti-worker, while public policy has been premised on the desirability of reducing wages.

I saw an interesting (and, I suspect, largely unconscious) illustration of this in a recent report from the grandly-titled Office of the Chief Economists. Among the many benefits of economic reform, the report cited the following

How does this relate to the wage share? When real wages are growing faster than GDP per person (and assuming a constant employment/population ratio, which is reasonably accurate), the wage share of GDP is rising. When real wages are growing more slowly than GDP per person, as they have done for the past 25 years or so, the wage share is falling. Looking at the beginning and end points we can see that wage growth has been slower than GDP growth over the period as a whole So, we can restate the conclusion as

Under the labour market institutions that prevailed between 1951 and 1981, the labour share of income increased. Since then, thanks to the adoption of market based approaches, workers have lost all the ground that they gained in the postwar decades, and then some.

30 thoughts on “Government report cheers for wage cuts

  1. Nothing to worry about – when the company tax cut goes through wages are going to rise – Messrs Turnbull, Corman and Morrison have said, so must be true.

  2. Yes but the workers, the older ones anyway, have gained because the value of their house has gone up a gagillion per cent. Then there’s their investment properties.

  3. re #2

    except the ones in comment 45 on the “Fortune favours the brave” 20th March column.

  4. Here we go again, “As market based approaches became more widespread, wages growth aligned more closely aligned more closely with growth in changes [sic] in GDP per capita.”

    ‘market based approach’ is identified as enterprise bargaining.

    The idea of a ‘market’ is to coordinate decisions made by agents in a society. The dreaded centralised wage fixing system did have a coordination role, albeit not perfectly in every conceivable respect. Enterprise bargaining achieved the opposite. So, there seems to be more than one notion of ‘a market’. In the former it is the welfare of all people, in the latter it is what?

  5. “Inequality is getting worse
    Trickle down? Our present era is one that is characterised by widening inequality where globally the richest 1 per cent own half of the world’s wealth — up from 42 per cent 10 years ago.”
    “JP Morgan and NAB argue the economic forces holding wage growth and better working conditions back are currently too strong, and the standard of living experienced by the majority of Australians now is as good as it is going to get for many years.”

    A man without a future is a man without a past. We have so corrupted our “system” that our young have no confidence in our politics, our economy, their own future in this country. Is this a good legacy, no it is not. My own 2 boys, young men, have taken their skills, their talent, their originality and their energy out of this country, I am soon to follow.

  6. You didn’t mention “while relying on growth in real wages to sustain the mirage of growth and good economic management.

    Every single Liberal budget talks about a wage breakout in two places: first as the main driver of the return to surplus that they’re forecasting; then later as a major risk that must be removed by further anti-worker action.

  7. It’s worth pointing out – hence the hint in the chart and the paragraph preceding the one you quoted – that much of the “ground lost” was in the decade beginning 1983 under that most decidedly non-market mechanism of the various (Prices and Income) Accords.

  8. Another case of lies, damn lies and selective statistics?

    Various complications and questions here:

    1. In 1951 a lot of labour wasnt actually included in the GDP stats in particular ‘women’s work’ such as child minding. In this case to varying degrees income came through the male ‘bread winner’ and was then dispensed out. If it had been allocated the income per capita would have dropped.

    2. Where does superannuation fit – is it counted as income?

    3. Is earnings before or after tax?

    4. The use of ‘average’ is a great way of hiding restribution upwards which could be more clearly identified by showing change between the median and the average but isnt. But this would be soooooo embarassing I suggest the constructors of the plot have avoided it.

    5. GDP doesnt cover improvement in product quality and lower item costs depending on you interest e.g. cars and cheap tools from Bunnings.

    6. Conversely is inflation adjusted GDP based on CPI an absolute measure. My reading here is its a lot more problematic than people realize

    7. There are other stats which might be a better indicator of life getting better or worse. One is life expectancy Despite that lack of improvement in wealth distribution we appear to have gone for example from 65 for males in 1951 to 80 today. This statistic is far more objective that (dodgy ?) than their economic cousins.

    a. Is this plot terribly reliable.
    b. Given how relatively close the start and endpoint are could it be that what we are seeing is a reflection of uncertainty and changes in the relative significance of the weighting factors.

    As is usual with economic graphs we cant actually tell, because to include uncertainty boundaries could reveal the scale of ignorance and simplification here. How you would do this I’m not sure as there is probably a lot of rubberiness in the indices which are drawn together into the simplification that is GDP let alone inflation adjusted GDP.

    In fact the narrative here may be correct – or things may be significantly worse or better depending on value judgement / perspective.

    For me though so far the stats picture does not pass the laugh test.

  9. @Mark Cully

    The point of the early Accords was to reduce real wages, because it was thought by everybody (whose opinions carried any weight) that they’d grown too much in the late 70s and early 80s. The reduced real wages were (it was said at the time) replaced by the “social wage”, a nebulous concept but which included the introduction of Medicare and other goodies.

  10. @Mark Cully

    I was just using the wording from the report. The main point of the post is that “public policy has been premised on the desirability of reducing wages. ”

    As I’ve said elsewhere, the current situation isn’t based on market forces either. The full force of the state is used against unions, while employers whose business model relies on wage theft get slaps on the wrist.

  11. The chart uses an old trick that is designed to mislead. It plots two variables on two different axes, but scales are different. So we see that $10,000 of GDP per capita on the LHS vertical axis is tbe same distance above the horizontal axis as $200 of real wages on the RHS.

    But $80000 on the LHS (a growth of 8 times) is the same distance above the horizontal as $1400 on tbe RHS ( a growth of 7 times). With this scaling, the two lines since the introduction of enterprise bargaining in 1993 track closely together, which is what was supposed to happen according to the narrative – real wages and productivity growing at the same rate.

    If the RHS scale had been drawn honestly with a real wage of $1600 at the same distance up as $80000 of GDP per capita, then the dark blue real wage line would be well below the light blue real wage per capita line. But this would not be convenient to the story by the Office of the Chief Economist.

    You can fool some of the people some of the time …

  12. So we see that $10,000 of GDP per capita on the LHS vertical axis is tbe same distance above the horizontal axis as $200 of real wages on the RHS.

    … It’s not, though. I dump it into illustrator and I find that the RHS scale marks are ~6.35mm apart and the LHS are ~5.64mm apart. To within the — poor — limits of accuracy, the LHS marks divide the space [~46 mm] into eight parts, the RHS into seven. Now that I look at it, it’s pretty clear by eyeball alone.

    “Yes, Collin, you are correct: my observation was wrong, I leaped to conclusions and I will be careful to avoid similar mistakes in future.” It’s not hard!

  13. If your eyes are bad and you don’t have my tools, btw, you could have checked this by scrolling so that one tick-mark is just about at the level of the bottom of your window, and the other one either clearly cut off or clearly visible above.

    Could have, but didn’t, because it didn’t fit with the story you thought you’d spotted.

  14. @Smith
    Since there are 7 tick marks and not 8 on the right scale, and the first tick mark on the right ($200) is not level with the first on the left, this is actually a non-point.

    As long as the scales are properly based to zero – and there is no evidence to suggest they’re not – then the scales are technically fine – just not as ‘neat’ as the human mind likes to see.

    To see this, the $700 mark on the right (50%) looks like it lines up perfectly with the $40k (50%) on the left.

  15. John Quiggin says: “while employers whose business model relies on wage theft get slaps on the wrist.”

    Amen to that. A worker who steals $50,000 from a boss will almost certainly have a conviction recorded against their name. On the other hand, if a boss does not pay a worker $50,000 in superannuation, award entitlements and the Fair Work National Employment Standard entitlements, it is a civil rather than criminal matter and the worker’s lawyer will usually urge the worker to settle out of court for a much lower amount because the costs of running a case are enormous. Even a cheap industrial law barrister will charge you $4k for a day in court and your solicitor will cost even more. Fair Work is also a “no costs” jurisdiction so the worker is left with their legal expenses even if he/she is obviously correct unless the boss’s litigation is deemed vexatious by the judge (in practice this almost never happens).

    Moreover, if your boss has set up a company, they can strip its assets and bankrupt it to avoid paying any court award. At most, he/she might get a small pecuniary penalty for accessorial liability that survives personal and business bankruptcy and the court can order that payable to the worker. To add insult to injury, the boss can then set up another company and do the same thing all over again. The name given to this rort is phoenixing. Some bosses are now on their fourth or fifth company, as I’ve discovered by reading the Federal Circuit Court industrial law cases on Austlli.

    I am only pursuing my own unpaid wages claim in the courts because I am now retired. If I hadn’t retired I just wouldn’t have the stamina or energy. And even if I do win, my lawyer says the legal fees may be as much as 50% of the court award. I’m not sure why I bother given the stress of litigation.

  16. @Collin Street

    The number of scale marks has got nothing to do with it. I could put 14 scale marks on the RHS showing real wages every $100 instead of every $200. The lines would not change either absolutely or relative to each other.

    But if my RHS axis goes from $0 to $1600, a more accurate representation, then the real wage line is well below the GDP per capita line.

    Or I could have the RHS scale up to $600 and say “wages have grown so much they’re off the chart!”

  17. @Smith

    You can download the data from the website and do your own analysis and make your own graph.

    See here for a more accurate representation:

    Between 1951 and 1993, employee earnings growth (2.0% per year) was essentially in alignment with growth in GDP per capita (1.9% per year).

    So, if you think that the wage share should be stable, we didn’t really need the period of “market based approaches” to achieve a realignment.

    Between 1993 and 2016, employee earnings growth was half that of GDP per capita (1.0% versus 1.9% per year).

    It’s most certainly not the case that “wages growth aligned more closely” with GDP per capita growth during this time.

    Cully makes one claim here and the exact opposite claim in the report. It’s bizarre.

  18. @Xevram
    I am glad to see that you “walk the talk”. That is the only way your post can be considered reliable. I tend to agree with you, except that my boys are still hoping to make it here somehow, in spite of the poor odds, so I’m hanging on. After all, they are all I have left.

  19. Peoples, you should know that Mark Cully is in fact the Chief Economist of the Department of /Industry, and hence the author (or at least the authoriser) of the report. Mark clearly assumed John knew that, but other commenters here may not.

  20. @derrida derider

    Office of tbe Chief Economist sounds very impressive … until you see it is in the Department of Industry. I’ll bet the economists in Treasury have a good laugh about it.

  21. The chart doesn’t tell us what you think it does. The period of high real wages coincided with a high unemployment rate.

  22. @derrida derider

    The report argues that the problem was that wages were increasing faster than GDP per capita from the 1950s to the early 1980s, causing inflation. It does not argue that wages were already too high relative to GDP per capita at the beginning of the 1950s.

    By the early 1990s, the increase in real wages since the beginning of the 1950s had been brought into line with GDP per capita growth, as my graph shows. Inflation was also low and fairly stable.

    Since this period, wages have fallen about 15% relative to GDP per capita.

    The report argues that “[a]s market based approaches became more widespread, wages growth aligned more closely with growth in changes in GDP per capita”. This is not an accurate or “neutral” description of the period from the early 1990s, irrespective of whether “aligned more closely with” means 1) wages fell relative to GDP per capita to achieve a realignment (this had already happened), or 2) wages grew in line with GDP per capita growth when before they grew too fast (wages fell relative to GDP per capita).

    I agree that the graph in your link includes more germane information, although it paints essentially the same picture.

  23. The number of scale marks has got nothing to do with it. I could put 14 scale marks on the RHS showing real wages every $100 instead of every $200. The lines would not change either absolutely or relative to each other.

    _Except_ that if you increase the number of ticks the lower tick will get closer to the zero mark. There are eight ticks to the left and seven ticks to the right; this means that the bottom mark on the left is closer to zero than the bottom mark on the right. As you can see.

    You claimed otherwise [you were wrong to do so] and then from this [false] claim derived another conclusion, that the graph was biased; because the conclusion is based on false claims the demonstration fails.

    This means that your claim was wrong, and that means you should withdraw it.

    But if my RHS axis goes from $0 to $1600, a more accurate representation, then the real wage line is well below the GDP per capita line.

    [what you write here is deeply confused, btw. You should give serious thought as to why you think this; it might help you see the nature of your mistake. I can help, a little, but it’s going to need a lot of work from your side]

    The top value on the RHS [and on the LHS] is arbitrary; any is as “accurate” as any other. The particular value was chosen so that the points were roughly equal on the _left_ hand side; this means that the percentage changes and change-rates — gradients — are effectively consistent/comparable across both sides.

  24. @Collin Street

    Yuo are writing garbage. Have a look at the graph in the document linked by derrida derider. Both variables are scaled as index numbers with the same base. This makes it so easy even you could understand it.

  25. OK.

    A question, Smith. In post #11, 23-march, you claimed:

    So we see that $10,000 of GDP per capita on the LHS vertical axis is tbe same distance above the horizontal axis as $200 of real wages on the RHS.

    Do you still believe this to be true?

  26. @Smith

    The indexed based chart is a better chart. Don’t think anyone would disagree with that. I even agree with you that the original chart arguably mutes the relative decline of real wages against productivity. But IMO that’s because the dark blue line starts higher, not because of scale trickery. The original chart is not objectively wrong.

    Let me put a challenge to you: If you were to accept the hypothesis that the original chart isn’t ‘tricky’ – but just not as clear as it could be – how would you go about defending it as technically accurate?

    Having said all that – I’m not even sure that there is any substantive disagreement here. We all agree that chart(s) show that real wages grew more slowly than real GDP per capita over the full period, and the ‘turning point’ was ~1983. Right?

  27. @EconoManOz

    The text says that since enterprise bargaining began in 1993 real wage growth “aligned more closely” with growth in GDP per capita. (The text mistakenly says growth in changes in GDP per capita). The chart by way of scale trickery shows the two lines close together after 1993 and far apart before 1993.

    This is not muting or mere lack of clarity. The indexed chart shows that post 1993 the gap is bigger and has lasted for longer than when the gap was the other way around. What the original chart tried to show was not an exaggerated version of what is true. It tried to show the opposite of what is true.

  28. @Smith

    Like I said previously, there’s no substantive disagreement here. The chart doesn’t back up the Chief Economist’s claim: real wage growth has been slower than growth in real GDP per capita over the whole period, or from ~1983 onward, or from 1993 onward – take your pick. I can see that in the original chart, or even easier in the index chart. One could plot growth rates directly and make it clearer still! I agree with Luke Elford at #17.

    Perhaps you’re right that they calibrated the chart to ‘try and show’ a particular story. I can’t prove they didn’t. But what’s the old saying about conspiracy versus stuff up? (I’m guessing, if you used that data and let the chart set the scales of the axes automatically, it would come up with those scales.) At least as plausible: they created that chart, and then misinterpreted the lines moving closer together.

    But you explicitly alleged scale ‘trickery’ – that the graph was technically wrong – because the first tick marks on each axis are ‘equal height’ and the right axis didn’t finish at $1600. That is objectively, provably, incorrect.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s