August 12, 2009

Life After People

Gregory Clark's column on the future of unemployment seems to have gone over like a lead balloon; both left and right have given it a good pummeling. To recap: Clark argues that continuing technological progress, particularly in computers, will make the unskilled portion of the labor force unemployable. As he puts it:
So, how do we operate a society in which a large share of the population is socially needy but economically redundant? There is only one answer. You tax the winners -- those with the still uniquely human skills, and those owning the capital and land -- to provide for the losers.
Now, there's certainly a kernel of truth to this: As Mike Konczal notes, the long-term unemployment rate is at an all-time high, and the trend seems to point to the continued disappearance of unskilled workers from the work force. But Clark seems to have closed off any possibility of our society adjusting to technological progress in the way that every generation since the onset of the Industrial Revolution has adjusted: Some jobs disappear, but new ones are created, and people train to fill those jobs. The process by which this happened, to be sure, was messy and often unjust, but one can't deny that each generation since the onset of the Industrial Revolution adapted in this manner; and that, outside of major financial crises, the developed world has not been afflicted with a large underclass of the permanently unemployed.

Induction, however, is a dangerous thing, and it's all too facile to say that the computer revolution (including the Internet, AI, and robots) won't be any more disruptive than previous technological breakthroughs. What can we say about Clark's argument that doesn't dismiss it as neo-Luddism? I find myself borrowing liberally from both Reihan Salam and Ryan Avent for my thoughts on this. Salam, echoing Will Wilkinson and Tim Worstall, points out that economic incentives for low-income households can affect enormously how technological progress impacts their livelihoods -- though I think that government tax and welfare policies have mattered less than automation and globalization have, in terms of mass unemployment among the unskilled.* I like Salam's idea of wage subsidies, so that jobs for the unskilled that pay less than what can be lived on can still be taken; still, he is right to note that financial remedies may not do much about the social disruptions that would take place in the wake of mass automation.

Avent, for his part, says that the dystopia of a robot-run society with billions of superfluous people is likely not even feasible, either physically or economically. That point is important: how our energy policy plays out over the next few decades will be a big determinant of how technological progress will affect low-skilled labor. I can see a world in which oil is over $200/bbl severely hampering the supply chain for computers and robots, along with everything else. Not that that would be insurmountable -- domestic manufacturing, including for robots, could well revive in that environment -- but paying attention to resource constraints is always important when talking about the future.** Too often, we get utopian happy talk that blithely assumes physical abundance will always be with us and that technology will overcome any scarcity issues. In other words, that man-made capital is substitutable for natural capital, rather than being complementary.

Avent also notes that Clark's focus on unskilled labor is bizarre given that computerization is likely to hurt the highly educated as much as, if not more than, the less educated. I'm currently reading A Farewell to Alms (albeit at a glacial pace), and it seems to be a theme of his that social change is the product of social, or even genetic, behavioral traits. The Industrial Revolution happened because Europe, England in particular, was populated with descendants of the rich, not the poor, and thus had the traits needed to build an industrial civilization. So the ability of institutions to improve the lot of the poor is fairly limited: Either you have the social environment for economic growth, something that develops slowly over long periods of time, if at all; or you don't. That may or may not be the case (I'm skeptical, to say the least), but transposing that analysis onto the present day economy, where workers of all backgrounds are in danger of being made redundant, seems to have caused Clark's argument to go astray.

* This points to an ambiguity in Clark's op-ed: Is he talking about just American workers or the world as a whole? The wage differential between the average American worker and the average Chinese worker, say, is pretty large, and I suspect it would take a while for even robots to surpass the Chinese in that respect.

** Of course, if we become less reliant on fossil fuels, unskilled labor may yet have a chance -- let's not forget the green jobs debate. Yet again, automation could make even that moot.

UPDATE: Also read Tom Lee's thoughts on the subject. For myself, if the robot revolution leads to a Star Trek-style socialist utopia, I think I could get on board.

No comments:

Post a Comment