Another 'automation is bad' argument....

I have just stumbled on a new article which is arguing that automation is not a good thing (see ref below and discussion of the article here) and this follows my discussion of Krugman's Luddite point of view some time ago. They discuss robots, but also quite reasonably discuss other forms of substitution for human labour, such as the ability for computers to calculate taxes, and automated toll collection etc. This new article argues that the rise of automation is going to cause an inter-generational wealth gap:

As shown below, in an admittedly highly stylized life-cycle model, the general equilibrium effects of this generational redistribution can transform enhancements in machines into very bad news not just for contemporaneous young generations, but for all future generations. The model treats all young workers as unskilled agents who invest their savings in the acquisition of both skills and machines. When today’s machines get smarter, today’s young workers get poorer and save less. This, in turn, limits their own investment in themselves and in machines. The knock-on effect here is that the economy ends up in all future periods with less human and physical capital, which further depresses the first-period wages of subsequent young generations. Although the skilled wage premium and the return to capital rises, the net impact of smartening up today’s machines is a reduction in the lifetime wellbeing of today’s and tomorrow’s new generations. In short, better machines can spell universal and permanent misery for our progeny unless the government uses generational policy to transform win-lose into win-win.

As a positive in the article, they recognise that the massive increase in labour in places like China and India has been a force that has driven down global wage rates. They also recognise that China is likely to also go down the road of automation, a point I made in my earlier post on the subject.
I can also at least give them credit for the novelty of their approach to the machines are bad argument.

However, the problem with their argument is very simple. It would apply equally to the Luddites, if not more. Indeed, today, capital is probably less concentrated than in 19th century England. So why is it that the industrial revolution created the many benefits that we still enjoy today?

After all, it was not just weaving in the industrial revolution that was impacted by automation, but many industries. In each case of automation, previously skilled work was automated, and also unskilled work, but the result was a revolution that saw long term improvements in standards of living. Modernity is the product of this revolution. The whole point of automation is to drive down costs of production, and this leads to cheaper goods, and cheaper goods means greater affordability, and more goods being available and so forth. Now if we jump back in time to the time of the Luddites, automation was undoubtedly a bad thing for skilled weavers, but also gave cheaper cloth for everyone else. This cloth was made in factories which themselves created a whole raft of new jobs, for example generating employment for skilled and/or semi-skilled engineers, as well as unskilled work in the manufacturers of the machines. As labour was freed as a result of industrialisation, new products could be made and the general stock of products available increased, and the cost of those products fell. In other words, automation destroyed jobs, but also created new jobs, and provided a wider range of products (and services) as lower costs.

There is no real difference here, except that some of the jobs that are being replaced are white collar. However, the idea that a computer can do complex tax calculations is directly equivalent to a skilled weaver doing complex weaving. The nature of the technology makes no difference, and nor does the status of who is being replaced. Indeed, if we read their introduction, we find this:

But what if the Luddites are now getting it right -- not for labor as a whole, but for unskilled labor whose wages are no longer keeping up with the average? Indeed, what if machines are getting so smart, thanks to their microprocessor brains, that they no longer need unskilled labor to operate?

Evidence of this is everywhere. Smart machines now collect our highway tolls, check us out at stores, take our blood pressure, massage our backs, give us directions, answer our phones, print our documents, transmit our messages, rock our babies, read our books, turn on our lights, shine our shoes, guard our homes, fly our planes, write our wills, teach our children, kill our enemies, and the list goes on.

Yes, technology has always been changing. But today’s change is substituting for, not complementing unskilled labor. Yesterday’s horse-drawn coaches were replaced by motorized taxis. But both required a human being with relatively little human-capital investment – a cabbie -- to drive them.

Tomorrow’s cars will drive themselves, picking us up, dropping us off, and returning home all based on a few keystrokes. This will make cabbies yet another profession of the past.

I have put emphasis on a couple of points. Take the example of the frivolous machine that massages our backs. This is a product which replaces a human masseur. That must be a bad thing, right? However, there is now employment in the manufacture of a new machine, and one which makes having an affordable massage available to more people. Automation has allowed this to be achieved, and for those (unlike me) who value a good cheap massage, their life is improved by this becoming affordable. Or take the idea of machines reading our books. For the blind, this constitutes (I guess) a huge potential improvement in their lives, and again, the development of software and manufacture of computers to do this is a source of employment. Sure, braille producers will eventually go out of business, but would we begrudge the availability of the books for the blind due to this loss of employment. In a recent example, Excel software undoubtedly would have led to less employment for the less skilled account clerks, but also would have created greater demand for those who could interpret accounts. Now, some of that interpretation is automated, and this in turn will reduce the need for skilled interpreters of accounts. In neither case did the employment world fall in as a result.

The important point is that in the introduction to the paper there is a whole list of new consumer goods and devices that were previously unimagined and that directly or indirectly result from the technology of automation. Who could have imagined a mobile phone 50 years ago (notwithstanding science fiction), but here they are. As technology improves further, other unimagined devices, services and consumer products of all kind will become available, whilst existing products will become cheaper and more widely available. New skills will be in demand. For example, as result of relative affluence in some countries, more people can afford to eat out, and this generates new demand for labour that is skilled, semi-skilled or unskilled. It allows for new leisure pursuits to be developed such as bungi-jumping. All of this is contingent upon the expansion of productive capacity, and automation does exactly this.

Just as the stock of types of goods increased and cost of goods fell in the industrial revolution, so the same will happen today, as exemplified in the examples the authors give. It really is no different. Where once there were skilled hand loom weavers, there were then other skilled and unskilled jobs that were created in their stead. This net increase in the availability of types of goods and reduction in the cost of producing goods can only be a good thing. However, automation of any kind sees a short term cost to those who are replaced.

The real difference today is nothing to do with automation, but is to do with something the authors themselves highlighted. There is indeed the question of the supply of labour in the world, and competition is particularly intense between unskilled labour. However, there is also intensifying competition in the middle as countries such as China seek to expand their higher education and increase the supply of graduates. This is nothing to do with automation, and everything to do with the supply of labour. As such, they are correct when they suggest that new Western workers face a challenging future, but are wrong about the reasons. I am genuinely puzzled at the revival of Luddism, and the rather odd justification for it. Is it simply fear of change? Comments welcomed.

Note: I realised I forgot a key point in the post just after hitting publish. What the authors are mistaken on is that they see automation as creating income inequality. What is actually taking place is greater income equality in which income is being redistributed to places like India and China, whilst income inequality within countries increases due to the oversupply of labour overall. Capital wins in a situation of an oversupply of labour as labour is unable to bargain for a greater share of profits. It really is very simple. Their claims of greater income inequality only stands up if looking within a country, not if looking between countries.


Sachs, J. D. and L. J. Kotlikoff (2012). Smart Machines and Long-Term Misery, National Bureau of Economic Research.

Krugman: The New Luddite

I know, I post too often on Krugman's musings in the New York Times. However, I could not resist the later missive from Krugman regarding automation (robots, and I presume he also means software that automates tasks currently performed by humans, but that is a guess):

If this is the wave of the future, it makes nonsense of just about all the conventional wisdom on reducing inequality. Better education won’t do much to reduce inequality if the big rewards simply go to those with the most assets. Creating an “opportunity society”, or whatever it is the likes of Paul Ryan etc. are selling this week, won’t do much if the most important asset you can have in life is, well, lots of assets inherited from your parents. And so on.

I think our eyes have been averted from the capital/labor dimension of inequality, for several reasons. It didn’t seem crucial back in the 1990s, and not enough people (me included!) have looked up to notice that things have changed. It has echoes of old-fashioned Marxism — which shouldn’t be a reason to ignore facts, but too often is. And it has really uncomfortable implications.

But I think we’d better start paying attention to those implications.

Welcome to the new Luddite approach to the improvement of the human condition. The robots will move wealth to the holders of capital from the workers.....and that is bad for us all. And what, might I ask, does he think happened in the industrial revolution? The hand loom weavers were put out of business due to automation. It was one of the heralds of the greatest changes in the history of mankind, and one from which we all now benefit. The industrial revolution was, in some respects, a painful process, but also served to create the potential for the huge benefits that we now enjoy.

In the modern and developed world, the adjustment to this new source of productivity should be less painful, albeit that the transition will not be easy either. However, unlike the industrial revolution, we are not transitioning from a system that was for many people barely above subsistence. Today, we are not barely above subsistence, but that is not to say that the transition will not be hard for some people. If you have spent years learning to do something like a search for legal precedents, the automation of this process will be a hard blow. However, for all of this pain, automation will have positive benefits alongside of these costs. In the case of legal services, access to the support of the law will be more affordable to those who would otherwise struggle to have access to the services, and that may matter a lot to many people. It is no different to the access to relatively cheap cloth that will have made new clothes more affordable to poorer people; at a cost to the hand loom weavers in the short term.
Just as the spinning jenny benefited the capitalists, the providers of automated legal searches will benefit the owners of law firms who adopt the technologies that are now available.

The examples of automation creating wider economic benefits are so many that it is difficult to know where to start. The end of canals in the face of the competition from more efficient trains (which created a boom and bust that left the UK with a railway system paid in full by the loss of the capital owners), the redundancy of the typing pool with the advent of the PC and so forth. In all these cases, automation has the result of getting more from less labour, and that means more 'stuff' per unit of employed labour, even in the case of the indirect effect of removing the typing pool. As fast as labour is being removed, more output is being created per worker, meaning that more 'stuff' is available per worker. With greater output of stuff per unit of labour, labour is freed to do new things. It is, just as happened in the industrial revolution, a situation in which new jobs arise to replace the old jobs, as people find new stuff to produce, as the emerging surplus of labour is absorbed in producing 'stuff' that was previously only afforded by the few. It is the process which generates greater wealth.

One wonders, when Krugman says that there are 'uncomfortable implications' and suggest that those implications 'need attention', what exactly he might mean? Does he mean that we should halt the current process of automation, perhaps smash the robots? I mean really, what does he mean? He leaves his article vague and open, hinting at ominous consequences. However, those consequences have precedents that (for once in economics) are clear. We see the result of the precedents all around us in the developed world; more than enough food, better health, heated and comfortable homes, our many forms of entertainments, our freedom to access information, and so the list goes on.

Think of the example of the impact of the PC on the typing pool. It saw what was a hard earned skill eventually made redundant (however, I taught myself touch typing, so not entirely redundant, but I do not need the level of skill of the typing pool where there is no 'backspace'). Should we have looked at the 'uncomfortable implications' of this shift? For a typist, it was undoubtedly not a good situation. However, would we turn back the clock, and stop this change if we could? We could certainly reverse the change, by making word processing software illegal, and blocking any web services that might offer a similar facility. In a few year time, with the magic of backspace button gone, we would have huge numbers employed in typing pools. Those same individuals will be drawn from the labour force, and will be an opportunity cost; the opportunity to do something else which is genuinely in more demand.

What we are discussing here is nothing more than a variant of the broken window fallacy; that breaking a window is a good thing as it creates employment. In breaking the window as a deliberate act, it creates employment, but employment with no real point. Better that the people employed in repairing the window are engaged in productive labour with a genuine demand, rather than an artificial 'created' demand. In the same way, better people are employed in new avenues than artificially supporting, or creating, employment through the rejection of more efficient means of engaging the same labour. It is a make-work scheme where there is no need for the work. Better that the labour is employed in creating real value.

If ever there were evidence that Krugman has nothing to offer, this rather odd article is the evidence. It is no wonder that he leaves the 'implications' and solutions unsaid. If he were to say clearly what his article implies, he would be ripped to pieces. It is, as the title of this post implies, nothing more than a disingenous revival of Luddism. Krugman hides in ambiguity, but the 'implications' are clear; break the machine to save the interest of labour. He cannot see that, painful as the adjustment might be, labour is also the recipient of the benefits of automation.

Further, Krugman writes from a US perspective, and the potential benefits to the US worker are obvious; the cost of labour differential is diminished through automation. Maybe the labour will no longer be the crude repetitive labour of yesteryear, but all those 'on-shored' factories that benefit from automation will nevertheless create new employment opportunities. There may, in other words, be losers from automation, but there will be many more winners; and the win will keep on delivering, just as the industrial revolution today reverberates to the benefit of all in our day-to-day lives.