Every now and then the “human labour will become obsolescent because of machines and technology” theme comes up, and I wanted to outline a few of the relevant arguments. Some people advance the theory in a more ‘doomsday’ way than others, and some people use this theory as a way to justify UBI (Universal Basic Income) or other variants of more-government. Among economists I’ve seen, they refer to this as Zero Marginal Productivity (ZMP).
The most prominent recent example showing the ZMP hypothesis would be the YouTube video, Humans Need Not Apply by CGP Grey. The video is fascinating, though I think it’s a little overblown and hyperbolic at points.
It looks like economists are split on this issue, even among the more free market leaning economists: Cowen, Kling, Caplan, Henderson, Boudreaux, and Eli Dourado.
Theory of comparative advantage
I think Don Boudreaux over at Cafe Hayek has the correct intuition about this idea. The ZMP hypothesis does not jive well with the theory of comparative advantage:
Even the lowest-skilled worker is capable of producing something of value. And the value of using that low-skilled worker to produce that something, rather than using some other means (i.e., a higher-skilled worker or a machine) to produce that something, rises the greater is the comparative advantage of those other means at producing something else.
So in other words, just because machines can do some/many things more productively than humans can – does not necessarily mean humans will go out of a job, it just means humans become relatively cheaper at doing those things they can do.
Substitution versus complementary
The idea here being that if you’re a low skilled worker, your job may be entirely substituted away by machines/technology. But if you’re a worker in a field that requires higher skill, or perhaps if your job requires some combination of technical skill and fluid intelligence, all this new technology will increase the value you can create. The trick is, we aren’t so good at identifying cases where this happens due to the ‘invisible hand’ of the market. This concept is mentioned in a recent EconTalk podcast episode: David Autor on the Future of Work and Polanyi’s Paradox. See David Autor’s paper here.
In the podcast Russ Roberts and David Autor comment that new technology and machinery will enable complements between human labour and machines/computers in ways we could never imagine. Another good point here is that machines don’t ‘own’ anything and so any increased productivity gained as a result of their use is not profit going to machines, it is increased surplus going to humans. Remember, when other people become more productive, we also benefit.
As industries get disrupted there will be jobs that go obsolete, and new jobs created elsewhere. Granted, the effects of this changed income distribution will not be even. Some people make an argument in favour of forcible government redistribution (e.g. UBI) because the disruptive effects of this technology won’t be evenly felt. Some people even go so far as to suggest that if left unchecked, it will create a dystopic future where those who own capital machinery/tech will be rich, and those who don’t own any productive capital will be poor and live a hard life. I really disagree with this specific vision, because the transition will take place over time – it won’t be instant. This will allow people to retrain or refocus their skills, and there may be new ways for them to use general skills or knowledge that they already have (the ‘complementary’ idea above). I think this type of UBI argument is a backdoor appeal ‘necessitating’ more government.
Besides, the net effect of all these new technologies will be to make goods and services cheaper. If the new capital/tech becomes really cheap, then surely poor people could just buy themselves some productive capital equipment and enrich themselves in this way. There’s also every possibility that we’ll use machinery or technology to augment our capabilities e.g. Exoskeletons, genetics, nanotechnology. See this example of an exoskeleton enabling a worker to become much ‘stronger’ and productive in ways that were not previously possible.
What do we do and where do we go from here?
I think labour market deregulation is essential, for reasons Isaac mentions in Keep Them Down, Keep Them Dependent. Nominal wages might fall, but real wages would rise. Reducing labour market friction would encourage employers to take a chance on new hires, rather than be overly cautious as a result of labour regulations.
Better education is required, to make sure that people are able to be productive in the more technologically advanced future. That said, I obviously disagree that this education has to be granted by the government, nor could the government effectively ‘pick winners’ in terms of directing students into the optimal fields. Governments (central planners) lack local knowledge, and there are big public choice problems (e.g. unicorn government thinking). For this reason, we should have more free market delivered education. This will help us use the price mechanism to ‘guide’ the next generation of students into studying fields that humanity actually desires, as well as guide people to retrain when technology obsoletes their current job. This would be vastly preferable to the alternative, which is politically determining what skills people learn.
I’m not 100% certain that the ZMP hypothesis won’t eventuate, but I’m reasonably optimistic that it won’t work out that way. I think the most likely scenario is that technological advancement will lead to less work and more abundance.
Image: New Scientist, Daewoo