Sunday, June 21, 2015

Will robots replace human labour and reduce real wage levels?


The potential impact of technological change on real wage levels in high-income countries seems to me to be more important than some other questions about the future that attract more public attention. In particular, the future of real wages must be much more important than income distribution, since there is not much evidence that income distribution has a significant impact on the well-being of the mass of the population. Real wage levels have traditionally determined how the mass of the population live their lives – how well they eat, the standard of housing they are able to afford, how much leisure they can afford, their ability to travel and so forth.

I made as similar point about the relative importance of real wages levels and income distribution last year in my review of Thomas Piketty’s book, Capital in the Twenty-First Century. I asked:
“What happens if technological progress makes capital a close substitute for labour? If a substantial component of the capital of the future can be thought of as a work-force of robots, the economic consequences might be a little bit like introducing slave labour to compete with the existing workforce. Real wages might fall under such a scenario, even though national income could be expected to continue to rise.”

I then went on to refer to an article on this blog a few years ago in which I asked: Will history judge Marx to have been right about the effects of technological progress on income distribution? Looking back now, I think the answer I gave was not too bad - but it was not particularly enlightening.

In trying to consider how to give a better answer I have been trying to come to terms with some maths in economic models relating to bias in technological change (including in a master’s thesis on technological change and capital labour substitution in Australian agriculture that I wrote over 40 years ago) and some relevant empirical research. I think some of this stuff is helping me to understand what might be going on, but in trying to explain it (even to myself) it is more useful to refer to some very simple economic models that can be described verbally.

The place I start is to consider what would happen if technological change consisted entirely of the introduction of robots that are very close substitutes for humans with respect to all attributes relevant to production processes. I then consider some implications of the deficiencies of that model.

As I wrote earlier, the consequences might be like introducing slave labour to compete with the existing workforce. Real wages might fall under such a scenario, but we should not be too hasty in reaching that conclusion.

It appears obvious that an increase in the supply of labour will cause the price to fall. People with rudimentary economics training might think of it in terms of the law of diminishing returns. As you add more labour, keeping other factors of production unchanged, the marginal productivity of labour tends to fall and this is accompanied by a fall in real wages. Of course, it is not even necessary to have a rudimentary knowledge of economics to grasp the idea that an influx of migrants which resulted in a substantial increase in supply of labour could reduce wage rates of workers.

The problem with that analysis is that it is unreasonable to expect other factors of production to remain unchanged in the face of an expansion in labour supply. An increase in quantity of labour will tend to raise the rate of return on capital (by raising the marginal productivity of capital) and thus provide an incentive for further investment. If the supply of capital is sufficiently elastic, real wages need not fall as a consequence of the increase in labour supply.

The potential for an expansion in labour supply to be consistent with higher living standards comes as no surprise to anyone familiar with empirical modelling of the economic effects of migration. For example, a recent study undertaken for Australia suggests that immigration has a strongly positive impact on labour participation, employment and wage levels.

So, in economic terms it seems that we would not have too much to worry about from an influx of robots who were just like humans.

However, the model of technological change I have presented above is deficient in several respects. One major deficiency is that technological progress consists of much more than introduction of machines that perform similar functions to humans. It also involves technical innovations that enable humans to do their jobs better and the introduction of superior consumer goods. If we take a broader view of technological change there is less room to fear that it might result in lower real wages.

Another major deficiency of the model is that it fails to recognize that the replacement of human labour by non-human labour is an ongoing process rather than a new phenomenon. Nick Rowe explained it this way a few years ago:
“Horses were once like robots. Horses could do a lot of the same work that humans could do. Humans and horses can pull things, if you feed them. But then mechanical horses, called tractors, were invented, that could pull heavier things with cheaper food. Tractors pushed horses' wages below subsistence, so the horse population declined.
The robot horse displaced horses, just as horses displaced humans from all the jobs where humans pulled things. But humans, unlike horses, can do lots of other jobs beside pulling things. Humans are very versatile. Horses can't really do anything except pull things. So humans switched to doing other jobs, while horses couldn't. And the marginal product of labour, and hence wages in those other jobs, increased. Horses and tractors were complementary factors to human labour in those other jobs.
But that won't happen if robots are invented that really are just like humans, and can do all the jobs that humans can do. Robots that are just like humans would be just like slaves, rather than like tractors and horses.”

What we are seeing now is robots that are displacing humans from a range of activities and freeing them to do things that robots can’t do - just as horses did. There are adjustment problems for people in the affected industries, but the impact on average real wages is likely to be positive. Over time, superior robots are likely to be invented that will replace the initial series of robots, just as tractors displaced horses. If robots can eventually reproduce like crazy, their capacity to live off “the smell of an oily rag” might mean that wages in many industries in which humans are currently employed will be driven below human subsistence levels.
 
However, it seems unlikely that robots will ever be viewed by humans as close substitutes for human labour with respect to all attributes relevant to all economic activities. My guess is that many humans will show a strong preference for some goods with a high human labour input e.g. home produced food, restaurant meals and beverages that are served by humans, live music by local musicians, handicrafts and works of art produced by humans, and some manufactured goods that individual humans have designed specifically for themselves or friends and relatives.


My bottom line is that over the next few decades the impact of robots in replacing human labour is likely to be a relatively small part of the total impact of technological change on the quality of life. Rather than worrying about robots replacing human labour perhaps we should be more concerned that the rate of technological progress may be slowing down. I will turn to that question in my next post.


Postscript:
I would like to draw attention to comments by Jim Belshaw (see below). Since the discussion may be of wider interest I will reproduce the main points here:

Jim:  Doesn't the evidence suggest that we have a higher proportion of the population employed in lower wage jobs and a higher proportion joining the ranks of the longer term unemployed? I accept that part of the impact is distributional and timing.

Winton: The evidence you refer to is one of the reasons I have been thinking about technological change and productivity growth. Some of the move to lower paid jobs and less job security could be associated with adjustment to technological change i.e. the timing problem you refer to. Some could also be associated with lower productivity growth and insufficient investment.
If tech change is a big factor I would expect it to be affecting older people, with young people finding it easy to pick up jobs created by new technologies. We do see older workers losing jobs, but we also see young people finding it more difficult to find employment.

Jim: We have seen lots of cases of older people losing jobs and dropping out of the work force. That was a particular feature of the early nineties adjustment. However, older workers are also more likely to be in "secure" jobs and to have been there for a time. There is a higher separation cost for the firm. This was a feature of Germany ten years back.
It is actually not clear to me how many jobs have been created by new technology compared to jobs lost. I am no Ned Ludd. I am well aware of previous cases (the industrial revolution is a huge example) where the application of new technology has produced long term gains. I would also agree and have been worried by what I perceive to be the slow-down in technological advance.
But we seem to be in a situation now where technological improvement is dominated by refinement, process improvement and cost reduction. I used to argue that we didn't need to worry about that because Government and community services broadly defined would redistribute benefits. Then and now there were just so many things that could be done to improve the quality of life.
I accept that was a naive view, partly because of globalisation, partly because of a cut-back in what Government might do. Realistically, the wealthier countries have to accept that they have reached a wealth peak, that competition will limit their gains while redistributing wealth to others.

Winton: I would have expected the cost reduction to have resulted in profitable investment opportunities and an accompanying expansion of employment opportunities. It doesn't seem to have happened and I don't really know why at this stage. I find it hard to perceive of cost reductions that do not increase profitability of investment. Perhaps we have reached the satiation point that Keynes wrote about, but I doubt it. 

Jim:  On Keynes, I doubt it too. I think one key issue with cost reductions lies in sustainability. There has been a problem with cost reductions designed to maximise immediate impact that have actually reduced value over the longer term.
There is also an issue that cost reductions increase the yield on what we do now but do not affect the yield on future investments. Increased profitability may increase the capacity to invest, but there is no necessary reason why additional investment should follow.
A recent RBA paper (referred to in a post on Jim’s blog) outlined the way in which investment decision processes (hurdle rates, pay back periods) might impede investment now. However, there is a timing issue here. If you accept that firm decision making processes have a degree of rationality, once firms are convinced that low inflation and lower interest rates will last for the immediate future, then the hurdle rate will come down.
The industrial revolution was based on the creation of mass markets. One of the difficulties in the thinning out of the middle class in many Western countries lies in the reduction of those markets. However, the mass market is growing elsewhere with economic development and globalisation. Investment rates in those countries are higher.

Winton: There are some interesting ideas there that are particularly relevant to Australia.
Another thought that has occurred to me is that a fair amount of the cost reduction is occurring in industries that are attempting to survive against competition from the free content available on the Internet. Think of the news media as an example. The internet is a major innovation providing substantial benefits, but causing a great disruption to the capitalist system as we once knew it. This is also part of the story about the apparent decline in the rate of productivity growth in the wealthy countries - the output of the Internet is not measured very well.

There has been a fair amount written around this topic but I have not yet come across anything that puts the pieces of the puzzle together in a coherent way.

18 comments:

  1. Not sure about this, Winton. Doesn't the evidence suggest that we have a higher proportion of the population employed in lower wage jobs, a higher proportion joining the ranks of the longer term unemployed? I accept that part of the impact is distributional and timing.

    ReplyDelete
  2. Hi Jim
    The evidence you refer to is one of the reasons I have been thinking about technological change and productivity growth. Some of the move to lower paid jobs and less job security could be associated with adjustment to technological change i.e. the timing problem you refer to. Some could also be associated with lower productivity growth and insufficient investment.
    If tech change is a big factor I would expect it to be affecting older people, with young people finding it easy to pick up jobs created by new technologies. We do see older workers losing jobs, but we also see young people finding it more difficult to find employment.
    There are also other factors involved e.g. quality of education.

    ReplyDelete
  3. Hi Winton. We have seen lots of cases of older people losing jobs and dropping out of the work force. That was a particular feature of the early nineties adjustment. However, older workers are also more likely to be in "secure" jobs and to have been there for a time. There is a higher separation cost for the firm. This was a feature of Germany ten years back.

    It is actually not clear to me how many jobs have been created by new technology compared to jobs lost. I am no Ned Ludd. I am well aware of pevious cases (the industrial revolution is a huge example) where the application of new technology has produced long term gains. I would also agree and have been worried by what I perceive to be the slow down in technological advance.

    But we seem to be in a situation now where technoloigical improvment is dominated by refinement, process improvment and cost reduction. I used to argue that we didn't need to worry about that because Government and community services broadly defined would redistrubute benefits. Then and now there were just so many things that could be done to improve the quality of life.

    I accept that was a naive view, partly because of globalisation, partly because of a cut-back in what Government might do. Realistically, the wealthier countries have to accept that they have reached a wealth peak, that competition will limit their gains while redistributing wealth to others.

    ReplyDelete
  4. Jim, I would have expected the cost reduction to have resulted in profitable investment opportunities and an accompanying expansion of employment opportunities. It doesn't seem to have happened and I don't really know why at this stage. I find it hard to perceive of cost reductions that do not increase profitability of investment. Perhaps we have reached the satiation point that Keynes wrote about, but I doubt it.

    There are a range of other possible explanations, but rather than rushing to judgement I would prefer to leave the question open and come back to it over the next week or so. Meanwhile I will bring your comments up into a postscript.

    ReplyDelete
  5. Hi Winton and thanks. On Keynes, I doubt it too. I think one key issue with cost reductions lies in sustainability. There has been a problem with cost reductions designed to maximise immediate impact that have actually reduced value over the longer term.

    There is also an issue that cost reductions increase the yield on what we do now but do not affect the yield on future investments. Increased profitability may increase the capacity to invest, but there is no necessary reason why additional investment should follow.

    Jut looking at things from the perspective of your line of discussion, the RBA paper outlined the way in which investment decision processes (hurdle rates, pay back periods) might impede investment now. However, there is a timing issue here.

    If you accept that firm decision making processes have a degree of rationality, once firms are convinced that low inflation and lower interest rates will last for the immediate future, then the hurdle rate will come down. There are signs of this happening in the cash awash super sector. Investment should then increase.

    The industrial revolution was based on the creation of mass markets. One of the difficulties in the thinning out of the middle class in many Western countries lies in the reduction of those markets. However, the mass market is growing elsewhere with economic development and globalisation. Investment rates in those countries are higher.

    Not sure how much sense all this makes. I'm just using your ideas to try to extend my own!

    ReplyDelete
  6. Jim,
    There are some interesting ideas there that are particularly relevant to Australia.
    A thought that occurred to me in the middle of the night is that a fair amount of the cost reduction is occurring in industries that are attempting to survive against competition from the free stuff available on the Internet. Think of the news media as an example.We have this major innovation providing substantial benefits but causing a great disruption to the capitalism as we knew it. This is also part of the story about the apparent decline in the rate of productivity growth in the wealthy countries - the output.of the Internet is not measured very well. There has been a fair amount written around this topic but I have not yet come across anything that puts the pieces of the puzzle together in a coherent way.

    ReplyDelete
  7. Slavery was considered a normal and good thing in the past until our moral have changed about the subject.

    A world with robot identical to human can be compared to slavery without moral implication. Which make me wonder : In the past, did the farmers or the general population raised against slavery purely based on the fact they were "stealing" their work ?

    ReplyDelete
  8. I would add that Milton Friedman (https://www.youtube.com/watch?v=3eyJIbSgdSE) pointed out that the concept of "immigrants stealing the population's work" came the first time because of the invention of welfare state.

    This seems to extend now to "robot stealing the population's work", even if welfare state does not redistribute wealth to robot.

    Why slavery was not, in the past, considered as bad based on the sole reason that it stole people's jobs ?

    ReplyDelete
  9. Hi Nicholas
    My understanding is that the main opposition to slavery arose because it was considered morally abhorrent to own another person. Perhaps you are right that there was also opposition from people who felt that slave labour was depressing wages.
    In the economics literature there has been discussion of whether slave ownership was becoming uneconomic in the U.S. prior to abolition. I expect this was associated with incentive problems. Unlike robots, slaves required supervision to ensure they earned their keep.

    ReplyDelete
  10. Um, Nicolas, the idea of immigrants steeling the population's work long pre-dates either Friedman or the welfare state.

    ReplyDelete
  11. Can you point out your historical sources about that ?

    ReplyDelete
  12. Hi Nicholas
    Regarding the history of opposition to slavery. Larry Siedentop's "Inventing the Individual", which I read recently, contains a good discussion of the role of Christian values in leading to opposition to slavery in Europe. This is a complicated and contested subject but very strong opposition was expressed by Gregory of Nyssa in the fourth century, based on ideals of equality.
    Opposition to the slave trade was led by evangelical protestants and Quakers. See: https://en.wikipedia.org/wiki/Slave_Trade_Act_1807
    Similar religious groups also played a big role in the abolitionist movement in the U.S. See:
    https://en.wikipedia.org/wiki/Abolitionism_in_the_United_States
    However, the politics of abolition was complicated. The idea that the Civil War was primarily about slavery seems to be wrong. The recent movie about Lincoln is probably not an ideal source of historical evidence, but the impression it gives that politics was complicated seems to be accurate. I wrote about it here: http://wintonbates.blogspot.com.au/2013/02/will-lincoln-encourage-people-to-give.html

    I can't find the article that suggested slave ownership was becoming uneconomic prior to the civil war. There is plenty of stuff relating to the question of whether the institution of slavery was holding back economic development - which is a different question. See: http://www.economist.com/blogs/freeexchange/2013/09/economic-history-2

    The trend in the real price of slaves actually seems to have been upward over the period from 1800 to the beginning of the civil war. See: https://eh.net/encyclopedia/slavery-in-the-united-states/

    ReplyDelete
  13. > Um, Nicolas, the idea of immigrants steeling the population's work long pre-dates either Friedman or the welfare state

    Any reference about your point ? Any instance were immigrant were considered stealing work when there was no welfare state ?

    > The trend in the real price of slaves actually seems to have been upward over the period from 1800 to the beginning of the civil war

    The argument about the price of slaves is different than people protesting that slaves were stealing their jobs. However, the argument about the price of slaves might also well apply for robots one day. (maintenance is not free)
    I suspect there is a relation between our vision of immigrants today because of our welfare state as Milton Friedman pointed out, and the fear of people about having robots stealing their work.
    It is really strange that no people seemed to complained in the past that slaves were stealing their jobs... but complains about robot.

    > Opposition to the slave trade was led by evangelical protestants and Quakers

    So it seems that the opposition was mostly based on moral principles instilled by religion. Those moral principles would not apply to robots.

    I'd like to point out a brillant interesting article from mises.org exactly on the topic : https://mises.org/library/let%E2%80%99s-hope-machines-take-our-jobs-we-want-wealth-not-jobs

    ReplyDelete
  14. Hi Nicholas. I would have to go back and trace, but I can give you an immediate example. Opposition to Chinese migration into the Australian colonies in the nineteenth century as based in part on fear that they would undercut Australian workers. Something similar happened in the lower Hunter in the early twentieth century, then Australia's great coal province, when the bosses brought in workers from the UK to undercut efforts by the local unions to try to improve working conditions that were very insecure.

    Freidman's point, was, I think, a little different. Leaving aside his interpretation of US economic history up to 19i4, I'm not sure that this is right, the creation of the welfare state did create a new cost dynamic, fear of immigration adding to tax burdens.

    ReplyDelete
  15. Thanks for the reference, Nicolas. His bottom line is similar to mine. Everything will be cheap except the stuff that remains relatively scarce i.e. land and human labour. The situation becomes more problematic if we think of the next 20 years or so, and add in international factor price equalisation (via trade and migration flows). I am currently wondering whether that leads to the unattractive scenario painted by Tyler Cowen in "Average is Over".

    My apologies also for mis-spelling your name!

    ReplyDelete
  16. No problem about misspelling, I had not even noticed.

    Jim thanks for those references I'll take a look.

    ReplyDelete
  17. hat is very nice & informative article keep it up....

    ReplyDelete