The trendy view that robots — or the ever increasing automation of human labour, at least — are going to cause major economic problems in the near future has started getting its backlash.
The argument is that we are entering a period when automation will very quickly replace huge numbers of jobs — some estimates say up to 70 per cent of existing American jobs won't exist by 2100 — and that that shift has already begun, explaining a number of concerning economic phenomena over the last 30 or so years, including the declining labour share of income, increasing inequality, and the decoupling of the median wage from GDP.
That analysis has led to some strange contortions from mainstream economists trying to conceive of capitalism in a world in which work was not necessary, leading some, myself included, to suggest that in that extreme example, it might be worth re-examining the basic tenets of economics.
But once the robots problem hit the mainstream, as evidenced by the Financial Times' Edward Luce writing that Obama must face the rise of the robots, it started being re-examined with a more critical eye.
The Atlantic's Derek Thompson argues that our problem now is "a deficit of demand", and our problem in the future can be dealt with in the future.
Matthew O'Brien, writing for the same publication, points out that what that deficit of demand means is that in the near term, automation won't lead to job losses, but it will keep pay well below where we'd like it. He concludes that "globalization, mechanization, and the decline of unions have all helped capital and hurt labor, but so has inadequate demand the past decade."
The TUC's Duncan Weldon has addressed the case of robots in the present day, and came to much the same conclusion. He writes that the rising profit share of income is concentrated almost entirely in the finance sector, and argues the likely cause is that that sector managed to ensure that the distribution of risk in innovation was spread widely, while the distribution of the the rewards was increasingly narrow.
Wheldon's conclusion is that the problem in the present day is less of a problem than it seems: with well-targeted redistribution of wealth, the benefits of productivity growth in the sectors where innovation has been successful can be used to pay for decent services everywhere else. What we're seeing is not, then, a crisis in automation, but a simpler crisis in distribution.
I am inclined to agree with Weldon when it comes to the present day. The effect of automation today isn't categorically different from from the effect thirty years ago, but it combines with the receding desire for redistribution and the slack demand stemming from the financial crisis with pernicious results.
But when it comes to the effects of future automation, no-one the attitude that "we'll deal with it when we come to it" strikes me as dangerous. We don't know a huge amount about what the effects will be, but it's clear they'll happen gradually, over the next century; there's the very real risk of a "boiled frog" problem, where we don't realise that the entire system is in crisis until its too late.
At best, if the predictions are accurate, we've got an upheaval of similar magnitude to the Industrial Revolution. That resulted in massive gains the world over, but only after well over a century of struggle. Life for the average factory worker in the 1800s was hardly better than it was for the average agricultural labourer in the 1750s, though you can be certain that the merchant class saw a hefty improvement. It took world wars, nationalised industries, continued worker's struggle and massive redistribution of wealth to temper the distortions down to a level which could be described as sustainable.
Ideally, we should be planning to achieve the gains of the Industrial Revolution without the 18-hour days, sundering of families and massive environmental upheaval that came alongside it. Burying our heads in the sand until the 21st century's dark satanic mills have already arisen is not the best way to bring that about.