Monday, 8 July 2013

The wastefulness of automation

My latest at Pieria:
"Chris Dillow observes that "one function of the welfare state is to ensure that capital gets a big supply of labour, by making eligibity for unemployment benefit conditional upon seeking work." And despite noting that when jobs are scarce, paying some to "lie fallow" so others can work might be a good thing, he concludes that "this is certainly not in the interests of capitalists, who want a large labour supply - a desire which is buttressed by the morality of reciprocal altruism and the work ethic." (emphasis mine). Basic Income, therefore, is not going to happen because capitalist interests, claiming the moral high ground, will ensure that it never gains political traction.
"But what if capitalists DON'T want a large labour supply? What if automation means that what capitalists really want is a very small, highly skilled workforce to control the robots that do all the work? What if paying people enough to live on simply is not cost-effective compared to the running costs of robots?  In short, what if the costs of automated production fall to virtually zero? 
"I don't think I am dreaming this. I've noted previously that forcing down labour costs is one of the ways in which firms avoid the up-front costs of automation. But as automation becomes cheaper, and the efficiency gains from automation become larger, we may reach a situation where employing the majority of people at wages on which they can afford to live simply is not worthwhile. Robots can produce far more for far less." 
The remainder of the article can be read here.


  1. Surely the hysteresis starts earlier with the inability of young people with potential from furthering their education. As a result we may fall short of the target level of automation, or at least it may take exponentially longer as we approach it with successively smaller pools of creative and educated generations who have the right incentive to be bothered to try.

  2. Who came up with that silly title? I thought you had lost your mind and turned into a luddite from the title, but then the article body is balanced enough. Is someone at Pieria responsible for the outrage?

    1. No, 'twas me. I could have called it "The limits of automation", I suppose, since the conclusion is that in a capitalist system automation is self-limiting.

    2. The main message I see in your post is rather that "A labour market that is skewed towards unskilled jobs when the workforce is more highly skilled and educated is malfunctioning." and you explain well why and that the luddite solutions are not valid.

      Getting out of that conundrum is the hard bit, in theory the badly or under employed people should just operate their own robots to produce the things they lack for themselves, and we need to understand why that doesn't happen, or too slowly.

      I don't think it's because of a lack of (financial) capital. To some extent saying we are in a "capitalist system" is another case of bad title. Capital plays only a very peripheral role in a modern market economy, whose core feature is being a distributed peer-to-peer decision making system.

      You can see that, for example, looking at the world's most economically successful people: hardly anybody gets rich because they just have capital, and indeed most successful people started with essentially no capital and even if they had access to it later, it was only a peripheral feature when it comes to explain their success. And it remains true as you go down through the distribution, the middle class is not middle class because it was endowed with a middling amount of capital!

      Also ideological anthropomorphism is not helping. "Capitalists", as a class, even if the label was not misleading, do not have systemic intentions beyond the micro-level individual decisions that happen where personal ambitions and preferences meet the game's rules. Economic happenings are the results of these myriad interactions between individuals and system rules, and is not anybody's grand plan. Saying that "capitalists want a large labour supply" is no more enlightening than saying that it rains because the clouds want to make us wet.

  3. I don't think it is as simple as this. There is an income effect. And there is the question of competition.

    Does automation favour profits over wages? I guess that depends on whether there are large enough economies of scale or other factors to create a natural global monopoly or cartel. Otherwise, competition will erode profit margins, leading to lower prices. Even if automation does increase firms' market power, the outcome for profits then also depends on whether anti-cartel policy can be implemented globally. More on that below.

    Does automation favour the elite egghead workforce over the masses? Again, it depends whether the skills (or contacts?) enjoyed by that workforce give them market power. I suspect the rapidly-advancing decoding of the human genome means that 50 years from now any highly prized genes will have become abundantly available within the population. I also suspect that, since we are now past 50% urbanisation globally, the bargaining power of lower skilled workers has reached an inflection point, and is on the rise now. Witness wage growth in China. In fact, I think the rise of automation will actually be driven by rising global labour costs.

    Even if automation does lead to much greater income inequality, it does not follow that unskilled workers will starve, or at least not in the longer term. To the extent that automation does have economies of scale, it needs a market in order to achieve scale. But if a large part of the workforce can no longer earn the disposable income needed, then the market for automaton-manufactured goods and services will whither. Having a large share of your workforce go hungry is also not politically sustainable. I guess it may ultimately come down to how much pain and disappointment people can take before they get angry and organised.

    1. I'm not sure you've really understood what I was saying.

      Firstly, my core premise is that increasing automation ERODES profits, because it creates a tendency to over-supply. Owners of production then prop up profits by forcing down wages for humans, thus cutting production costs. This process would be exacerbated by competition.

      I see no evidence that the bargaining power of lower skilled workers is increasing, and considerable evidence that it is actually falling. I agree with you that rising labour costs drive automation - that incidentally is Marxian economics. However, one untoward effect of this is that medium-skill jobs are more likely to be automated than lower-skill ones, because the costs of automating lower-skill jobs exceed the remuneration of the workers. Forcing down wages is a way of avoiding the up-front cost of automation. Can I suggest you read the linked article "The financialisation of labour" that explores this more?

      Your third point exactly duplicates what I said in the article.

    2. I think one of the things that is of major influence is that medium skill jobs may actually be easier to automate than low skill jobs. Low skill jobs are things that anyone can do, say things like picking a piece of fruit off a conveyor belt that doesn't have the right Aesthetic qualities for Supermarket X and dropping it into a hopper of rejects. It's a skill that is easy for humans working on instinct, but relatively hard to program into a computer.

      Middle class jobs are skills that often act along simple rule based schemes, but it's the size of these schemes that is the thing that requires a large amount of time investment by the learner to commit to memory.

  4. It would not do to criticise on grounds of originality,but many of your ideas about automation and the consequent lack of effective demand were covered by Major C. H.Douglas from 1920 onwards, having an enormous influence in the 1930's when things came unstuck last time. His solution was bank nationalisation and the issue of unearned incomes for all called National Dividends. These were paid by gearing the creation of money mechanism to this end.(There was a Just Price mechanism to prevent inflation I believe, which I am a bit hazy about) A popular variation of his scheme was the Discounted Price mechanism where goods were sold at cost, then the tickets of sold goods were remitted to the Government who paid the manufacturers the profit out of the same ab nihilo mone y.Douglas often spoke on the radio and the Greenshirts marched demanding Douglasite reform,although the Social Credit government in Alberta actually enacted Silvio Gesell's velocity money instead,not issuing National Dividends but speeding up the velocity of money in circulation to a colossal degree.This scheme was echoed by a successful Gesellian experiment in the Tyrol where the Chiemgauer is a present day renewal of that economic philosophy.Keynes was influenced by both Douglas and Gesell,but opted for the velocity of circulation stimulus of Gesell minus his careful attention to the problem of increasing land values which hampers Keynesian demand management now: demand stimulus just inflates housing bubbles.

    1. I'm certainly not claiming originality. Much of what I cover in this post was foreseen by Marx, Kalecki, Keynes and others. Gesell I have come across but not Douglas, so thanks for updating me.

    2. And I was not claiming you were being unoriginal! Originality cannot be the issue in a discussion of the fundamentals .For what its worth I prefer Gesell or Keynes if e latter had included some safeguarding to stop monetary stimulus being hedged into land values the way Gesell did .(There is a log passage on or encomium of Gesell in Keynes General Theory with disapprobation for land value side of his work which is precisely where Keynes went wrong) Also as somebody else has said here, there is nothing stopping modern governments spending new money into existence and increasing demand by boosting the public sector. Or dare I say it reviving nationalisation (banking ,pubs!)

    3. Hehe. I've already suggested nationalising banking!

  5. Since in the Western economies, the service sector provides most of the jobs, I wonder why people speak of the consequences of automation as if they were something to happen in the future. The shift away from manufacturing has already happened! It's true that now we are seeing automation of some service jobs, but we know that not all service jobs are automatable. History will simply repeat itself: Less and less jobs in automatable tasks, and more jobs will appear of the non-automatable kind. As long as machines and the electricity/fuel they need to run are cheaper than paying people (and this may change in the future), we'll keep seeing this trend. We could, in principle, pay people a citizen's salary instead of employing them in things that, now that we come to think about them, aren't all that necessary. But that would probably change things too much for comfort for too many people, so we don't.

  6. This was one of the things that got me thinking about the politico-economic model that Labour ran up until the banking crisis - ie that you tax some of the excesses of capitalism to pay for the negative externalities and inequalities caused by it. (For example through tax credits). That model has since been shot to pieces both by the banking collapse and international tax arbitrage.

    There are a whole host of challenges that are then raised. One of them is about the geo-political structures of firms and regulators - and how to make both democratically accountable and transparent. The problem is that no single policy alone will be the magic wand. Calling for maximum ratios between the highest & lowest earners, or clamping down on tax avoidance alone will have limited impact. As we've seen with banking, the lobbyists will turn their firepower to make such rules full of holes.

    At the super-wealthy end, they have more money than they can spend - literally. You only need to scan the online tabloids to get an idea of the latest conspicuous status symbol. At the other end, you've got people wanting to do good things in their communities but unable to make ends meet if they did them. The number of otherwise very talented people I've met who've said they are 'selling out for the corporate dollar' is depressing.

    At an even greater 'big picture' level, are 'Western' levels of consumption sustainable? Not just from an environmental perspective but also from a simple 'making ends meet' level. Especially given rising global demand for price inelastic commodities.

    One of the mistakes the UK has made is 'relying on the market' for things that actually require longer term strategic planning - like transport and energy infrastructure. I'm thinking in particular where the sunk costs are the things the market doesn't like to provide for alone. Crossrail in London is a good example.

    I'm rambling so I'll stop

  7. Enjoy your suggestions! Cheers a lot with regard to sharing! I merely joined writer and started my own, personal website:) The very last images will be epic! Thanks thank you we appreciate you discussing this kind of content blogposts! Only found your site and i am actually astounded by your creativity :Deborah it's extremely neat whatever you accomplish!
    Blade & Soul Gold
    Diablo 3 Gold
    runescape gold

  8. Quality furniture and toys at unbelievable prices! Kidkraft is one of our most popular brands, which provides the best furniture for children's rooms and the whole house. We also provide quality toys, such as kitchen play sets, railroad train tracks, rocking horses, and more! We compliment our line of toys with AZ Importer who supplies top of the line radio controlled water boats, racing cars, monster trucks, and amazing helicopters. All of our products are proven in the marketplace by other consumers who have been very pleased with all of their purchases from us. We are adding new products daily, so check back often to see what is new today!remote control car.

  9. Where will the customers come from?

  10. Banking is highly automated, am I right?

    No, pen and paper ledgers, Less paper drafts.


    Electronics manufacturing jobs are 1/2 since the dot com bust! That sector I think has had one of the largest drops percentage wise.

    Make your own comparisons ticking off the series you want to see here:
    (Subtotals are above indented series.)

  12. Using logscale option in graph type setting lets you make a good comparison.

    Slope is the growth rate in logscale graphs.

    Globalized and automated: