Inequality

[Neil]

Could the collapse of organized labour as a meaningful force be a factor?  The timeline seems to fit.

[grumbler]

Could the collapse of organized labour as a meaningful force be a factor?  The timeline seems to fit.

I'd think it was a consequence, not a cause, but I could be wrong.

[The Minsky Moment]

My objection is that higher capital costs are ignored in the calculation, or at least are not shown to be irrelevant.  Now that you have an E Machine, you've got to take care of both capital and wage costs, where as previously you only had wage costs.  The labor cost of the bushel may have gone down from 10 cents to 5 cents, but the capital cost of the bushel had to have gone up by x cents, where x > 0.

Sure it's a cost, but it's not ignored.  The machine isn't used unless net it is profitable.  And to the exent the costs decrease the Farmer-Capitalist returns, they are increasing those of the E-Machine producers and workers.

[DGuller]

My objection is that higher capital costs are ignored in the calculation, or at least are not shown to be irrelevant.  Now that you have an E Machine, you've got to take care of both capital and wage costs, where as previously you only had wage costs.  The labor cost of the bushel may have gone down from 10 cents to 5 cents, but the capital cost of the bushel had to have gone up by x cents, where x > 0.

Sure it's a cost, but it's not ignored.  The machine isn't used unless net it is profitable.  And to the exent the costs decrease the Farmer-Capitalist returns, they are increasing those of the E-Machine producers and workers.

Yes, but not linearly, which is the trend that was assumed to be normal.  If I replace 5 cents of labor costs with 4 cents of capital costs, then I can only bid up the wages by 20% to $1.20 and not $2 (assuming the same margin), whereas BLS statistics would show 100% productivity improvement.  Maybe it doesn't explain why this 20% is about 0% in practice, but it explains a great deal why it's not 100%. 

[crazy canuck]

Could the collapse of organized labour as a meaningful force be a factor?  The timeline seems to fit.

I'd think it was a consequence, not a cause, but I could be wrong.

I agree, I think it was a consquence of workers moving away from jobs that were traditional trade union jobs into areas unions were less successful in organizing - ie higher skilled positions created by technological advancements. 

[The Minsky Moment]

I was always under the impression that the divergence between productivity and compensation was explained by exactly the factor that DG brought up:  that productivity is increasingly a product of investment in workers and in capital improvements, rather than labor investments per se. 

If that were so, then divergence should have been constantly incresing since the Industrial Revolution, which clear is not so.  That is the Marx thesis.

Another way to think about it is to apply to the facts re divergence in my first post.  In the theory is true, then it most be true that investment in worker human capital, physical capital and technological efficiencies must have been far greater in the period after around 1980 (when divergence occurred) then the period rom 1947-1982.  Indeed, since no divergence occurred in 1947-1982, one would expect to see stagnation in investment in human and physical capital and technology.  But that is not a very plausible claim to put it lightly.  That time period saw probably the two most efficiency-driving inventions of the century: containerization and the Green Revolution.  It also saw the GI Bill and huge increases in the proportion of workers receiving higher education.  In terms of physical capital investment, it is roughly equal throughout both periods.

[DGuller]

My objection is that higher capital costs are ignored in the calculation, or at least are not shown to be irrelevant.  Now that you have an E Machine, you've got to take care of both capital and wage costs, where as previously you only had wage costs.  The labor cost of the bushel may have gone down from 10 cents to 5 cents, but the capital cost of the bushel had to have gone up by x cents, where x > 0.

Sure it's a cost, but it's not ignored.  The machine isn't used unless net it is profitable.  And to the exent the costs decrease the Farmer-Capitalist returns, they are increasing those of the E-Machine producers and workers.

Yes, but not linearly, which is the trend that was assumed to be normal.  If I replace 5 cents of labor costs with 4 cents of capital costs, then I can only bid up the wages by 20% to $1.20 and not $2 (assuming the same margin), whereas BLS statistics would show 100% productivity improvement.  Maybe it doesn't explain why this 20% is about 0% in practice, but it explains a great deal why it's not 100%.

Actually, speaking of the median wage growth being about 0%, is median even a good statistic in that sense?  Median growth would be appropriate to analyze if the distribution of wages didn't change.  In that case, you want to use the median, because average can be distorted by really high values.  But, if the distribution of wages does change, which is kinda tautologically true given the thread title, wouldn't median growth be misleading?  Maybe that $0.20 per bushel that you saved with the E Machine now went to the newly created fat cat position of CFO to oversee the capital spending on the tractor, whereas previously you didn't need a CFO as you didn't have a lot of capital costs.  That fat cat's salary growth would barely be reflected in the median measure.

[DGuller]

I was always under the impression that the divergence between productivity and compensation was explained by exactly the factor that DG brought up:  that productivity is increasingly a product of investment in workers and in capital improvements, rather than labor investments per se. 

If that were so, then divergence should have been constantly incresing since the Industrial Revolution, which clear is not so.  That is the Marx thesis.

Another way to think about it is to apply to the facts re divergence in my first post.  In the theory is true, then it most be true that investment in worker human capital, physical capital and technological efficiencies must have been far greater in the period after around 1980 (when divergence occurred) then the period rom 1947-1982.  Indeed, since no divergence occurred in 1947-1982, one would expect to see stagnation in investment in human and physical capital and technology.  But that is not a very plausible claim to put it lightly.  That time period saw probably the two most efficiency-driving inventions of the century: containerization and the Green Revolution.  It also saw the GI Bill and huge increases in the proportion of workers receiving higher education.  In terms of physical capital investment, it is roughly equal throughout both periods.

I guess there are also multiple types of productivity gains.  Some gains of efficiency are just free lunch, such as those that come from scientific breakthroughs or just plain genius ideas that no one thought of before.  Other gains of efficiency are achieved only if you sink a lot of capital to get them.  I imagine that those two kinds of productivity gains have different implications.

[Jacob]

Could the collapse of organized labour as a meaningful force be a factor?  The timeline seems to fit.

That'd be one of my guesses, but I don't think it explains everything.

Interesting thread.

[Jacob]

But, if the distribution of wages does change, which is kinda tautologically true given the thread title, wouldn't median growth be misleading?  Maybe that $0.20 per bushel that you saved with the E Machine now went to the newly created fat cat position of CFO to oversee the capital spending on the tractor, whereas previously you didn't need a CFO as you didn't have a lot of capital costs.  That fat cat's salary growth would barely be reflected in the median measure.

That would've been my other guess.

[The Minsky Moment]

Actually, speaking of the median wage growth being about 0%, is median even a good statistic in that sense?  Median growth would be appropriate to analyze if the distribution of wages didn't change.  In that case, you want to use the median, because average can be distorted by really high values.  But, if the distribution of wages does change, which is kinda tautologically true given the thread title, wouldn't median growth be misleading?  Maybe that $0.20 per bushel that you saved with the E Machine now went to the newly created fat cat position of CFO to oversee the capital spending on the tractor, whereas previously you didn't need a CFO as you didn't have a lot of capital costs.  That fat cat's salary growth would barely be reflected in the median measure.

You are on to something here.  There is a divergence between median and mean compensation.  As it turns out, that divergence accounts for over 40% of the divergence of the observed divergence between median wage growth and productivity since 1973.  What is being captured here is increasing inequality in income.  What it doesn't tell us is why or how it is happening.  Why do these fat cats suddently spring up in the early 80s and why do they keep getting fatter?  And why are the thin cats failing to pick up any scraps?

[DGuller]

Actually, speaking of the median wage growth being about 0%, is median even a good statistic in that sense?  Median growth would be appropriate to analyze if the distribution of wages didn't change.  In that case, you want to use the median, because average can be distorted by really high values.  But, if the distribution of wages does change, which is kinda tautologically true given the thread title, wouldn't median growth be misleading?  Maybe that $0.20 per bushel that you saved with the E Machine now went to the newly created fat cat position of CFO to oversee the capital spending on the tractor, whereas previously you didn't need a CFO as you didn't have a lot of capital costs.  That fat cat's salary growth would barely be reflected in the median measure.

You are on to something here.  There is a divergence between median and mean compensation.  As it turns out, that divergence accounts for over 40% of the divergence of the observed divergence between median wage growth and productivity since 1973.  What is being captured here is increasing inequality in income.  What it doesn't tell us is why or how it is happening.  Why do these fat cats suddently spring up in the early 80s and why do they keep getting fatter?  And why are the thin cats failing to pick up any scraps?

One reason could be that financial deregulation increased the ability of the financier fat cats to actually add value, which increased both their numbers and their value add, which in turn bid up their wages.  EDIT:  IIRC, productivity growth rate itself increased at about the same time as median wages decoupled from productivity growth, so maybe the fat cats were both the main drivers and the main beneficiaries of that trend.

[The Minsky Moment]

To be clear, for this exercise I am essentially taking as a starting point that the neo-classical axioms hold and neo-classical theory holds, so that there is perfect competition, rational agents, etc. and Say's Law strictly holds.  The next step is to allow relaxation of those assumptions but specify the conditions that an alternative theory must satisfy - namely that it must explain in the US context why no divergence occurs from 1947-1982 but then big and increasing divergence occurs afterwards.

[Martinus]

Actually, speaking of the median wage growth being about 0%, is median even a good statistic in that sense?  Median growth would be appropriate to analyze if the distribution of wages didn't change.  In that case, you want to use the median, because average can be distorted by really high values.  But, if the distribution of wages does change, which is kinda tautologically true given the thread title, wouldn't median growth be misleading?  Maybe that $0.20 per bushel that you saved with the E Machine now went to the newly created fat cat position of CFO to oversee the capital spending on the tractor, whereas previously you didn't need a CFO as you didn't have a lot of capital costs.  That fat cat's salary growth would barely be reflected in the median measure.

You are on to something here.  There is a divergence between median and mean compensation.  As it turns out, that divergence accounts for over 40% of the divergence of the observed divergence between median wage growth and productivity since 1973.  What is being captured here is increasing inequality in income.  What it doesn't tell us is why or how it is happening.  Why do these fat cats suddently spring up in the early 80s and why do they keep getting fatter?  And why are the thin cats failing to pick up any scraps?

Wouldn't automatization be the explanation here too? You need a smaller cadre of highly skilled workers - everyone else is unskilled labour. This explains inequality.

Notice that this phenomenon is not just limited to industry - "soft" businesses, like law, experience this too. Compare the efficiencies of running a law firm with e-mail available, and only with courier/fax. We are now outsourcing stuff like due diligence reviews to Delhi - this means we will still need less highly skilled, highly paid lawyers in future.

We are reaching a paradoxical society where the rich elite is actually working for the poor masses - and the redistribution is the only thing keeping them from raising in a revolt.

[crazy canuck]

And we were doing so well.