The NIC team labeled the blog subject for this week: How Will Employment Change with the Expansion of New Technologies—like Robotics—in Manufacturing? Will We See a New Unemployable Underclass?
I’m hoping for a vibrant discussion with good back-and-forth among commentators this week. I’ve noticed that the blog up to this point has been populated by slightly longer, carefully crafted, ‘point of view’ statements — all of which are quite interesting — but a bit less interactivity than I would have expected. People will do on a blog what they wish to do… but let me suggest that it might be an interesting change of pace, and a good one, for commentators to feel that the barrier to entry is a little lower. So, this week, please feel free to throw out short comments, half to three-quarter baked ideas, questions that you’d like to see others take a stab at, etc. etc. Blog posts don’t need to be fully articulated arguments with careful language; that’s why they call it a blog. If you have a thought worth sharing, please share it and let others build on it.
I’m particularly excited to be moderating this discussion because I think it is a subject where the blog can really help to press the GT2030 draft document a step or two forward. I know my friends in the writing team will understand the motivation with which I make this comment: I think the question needs to be expanded and made much more ambitious.
I mean that in several respects, which I’ll suggest as possible directions for elaboration that I hope might provoke others to comment.
When it comes to manufacturing, robotics is already here, employment is becoming a different thing than it was, there is more or less a permanent underclass…. and it is a long way from 2030. If you’ve been in an auto factory, a steel mill, or even a semi-conductor fab in the last few years, you know what I’m talking about.
This is one of those (frequent) times when talking about the future, helps us to catch up to the present. ‘Manufacturing’ can mean different things to different people, but if the image of a ‘factory’ economy comes into your mind, think again. Emphasizing well-known numbers here: in the US manufacturing is well below 15% of GDP (and the recent spate of news about ‘insourcing’ back to the US is a tiny phenomenon in perspective, not big enough to move this needle). In Germany manufacturing is around 20% of GDP; in China, the ‘factory’ of the world, it’s around 1/3.
For the world as a whole, manufacturing as % of GDP is around 20%. For one fifth of the economic activity in the world, it sure does attract a lot of attention.
We’d need a new, different, and surprising story about the nature of economic growth and change over the next decade, to believe that any of those numbers are going to rise meaningfully, rather than continue to shrink on aggregate as they mostly have been doing for decades now. Manufacturing right now still has a surprising hold on (some) peoples’ imaginations, but it is hard at least for me to imagine that this will continue for another 20 years.
Robotics, or maybe we should simply say ‘automation’ to make it sound less exotic, is certainly part of the story of the employment trap. The productivity of what we call a factory has, in most modern sectors, skyrocketed over the last decade — in part because of automation, in part because of management paradigms, in part because of instrumentation and sensorization… and none of those trends is slowing down. It maybe that impact of sensors and data — measuring what is really happening at a granular level in the manufacturing process and using that to drive constant, incremental improvements — will be more impactful over the next decade or so than will robotics. After all, computation and data visualization is (right now) easier than robotics and there is a lot of inefficiency to be taken out of most manufacturing processes simply by understanding them better and removing waste.
What about employment? Could manufacturing employment approach zero, (asymptotically) by 2030, in rich countries? In medium income countries? Possibly even in poor countries? And since the really big benefits to productivity are to be found by incorporating (even fairly simple) ‘robotics’ or at least automation into services, how quickly do we think the question will shift in that direction?
Finally (for now) consider the notion of a ‘permanent underclass’. Is that 2030, or is it right now? The proportion of US GDP that goes to labor is at or near an all time low. Low-skilled manufacturing jobs that haven’t been lost to China, and lost from China to lower-wage locations in East Asia (and probably over the next decade to in turn to Sub-Saharan Africa), already constitute the employment contract for a permanent underclass in the US. Try living on $13.20 an hour. More important, try pulling yourself or your children out of the underclass with that wage as a foundation. If you think it will get easier over the next decade to do that, we need an explanation as to why.
A last provocation. It is conventional wisdom, and correct, to say that your chances of being stuck in this permanent underclass are inversely related to your level of education, at least in the US. That’s the good news. The bad news is, the return on investment in higher education — while still positive — has been declining now for 20 years. (Some of that is attributable to the massive rise in the costs of higher education, and some attributable to the increased competition from larger numbers in the global higher education pool — if anyone has seen a plausible analysis that breaks down those components, please share it). It seems almost inevitable (at least to me) that the costs of higher education will shrink and the accessibility of it multiply over the next decade as simple e-learning technologies take hold — and that’s just the obvious disruption we can already see coming. That means a much larger global class of educated workers looking for jobs. Where will the threshold for ‘underclass’ be then?
In the long run, there need be no negative trade-off between productivity and employment of course. But the long-run (in a theoretical economics sense) is probably not the relevant time frame here. Right now, many people are acting as if they believe that a job gained in one place is necessarily a job lost somewhere else. Jobs mercantilism, if you will. What are the plausible trajectories out of that dilemma?
Let’s not shy away, in the end, from speculating on what we might end up meaning by the term ‘employment’ — what it implies, for whom, on what terms, with what economic (and broader) impact on human life and meaning.
Some provocations, then, to start. Please have at it.