Jobs, robots and self-driving vehicles
Lately I’ve been reading Tim Dunlop’s excellent book Why the future is workless , and thinking about the issues it raises, particularly in the light of the prospect of autonomous vehicles and other transport technologies. Tim raises the obvious question: what will happen to people who currently drive for a living, and the broader issue of whether any kind of work will survive the process of automation.
Dealing with the second question first, labour-saving technological progress (aka automation) has been going on since the Industrial Revolution of the late 18th century. Until recently, it seemed (to me, at any rate) as if the process was decelerating, with technological progress being confined to information and communications technology. But, as I said in my previous post, the technological revolution seems, quite suddenly, to be affecting the transport sector. Looking at recent progress, I’d say that, within a decade or so, autonomous vehicles will probably be as cheap as human-driven ones, and safer than the average human driver, at least in for the kinds of travel typically associated with transport sector jobs.
There’s plenty of room for dispute about estimates like this. But, even if I’m right, that doesn’t mean that transport employment is going to disappear in the next ten years. It would probably take another ten years for sales of human-driven vehicles to drop to low levels, and another ten or twenty for them to disappear from the fleet.
In the meantime, technological change in the form of online shopping is likely to generate increased employment in goods delivery. The fact that, nearly 15 years after it became a major phenomenon, online shopping has not prevented growth in retail employment is itself an indication of how slowly these processes move.
Still, once the transition to autonomous vehicles starts, I’d expect to see transport employment dropping by 2 to 3 per cent a year. Using the Rule of 72, that would imply a reduction of 50 per cent over 24 to 36 years. And, if the tech revolution spreads more broadly, this could happen in sectors other than transport. Even so, this won’t happen evenly, so the rate at which jobs are displaced will be less than the rate in sectors where the transition is taking palce.
There’s nothing unprecedented in this rate of change, and no necessary reason why it should produce social dislocation. That doesn’t mean that we should accept the Pollyanna view that everything will be rosy. A look at the history of the past 250 years shows plenty of examples where automation has led to widespread distress and immiseration, set against a general trend of increasing living standards. Unemployment has fluctuated, without any obvious trend or apparent correlation with particular technological innovations.
In this context, the past few decades have been among the relatively bad periods, particularly in the leading industrial economy, the US. Wages have stagnated and the ratio of employment to population has fallen, particularly for men, but more recently for women also.
There are lots of things going on here: the breakdown of the Keynesian boom and the rise of financial capitalism being the most obvious at the macro level. But in terms of labour markets, the really striking thing has been the end of the long trend towards reductions in weekly and annual working hours for full-time workers. That’s evident globally, but more easily seen in Australia where standard hours are regulated by the Fair Work Commission and other bodies.
Over the 100 years or so to the early 1980s, Australian workers experienced steady reductions in annual and lifetime hours of work. The standard working week was first capped at 48 hours (an Australian first, celebrated for many years as Eight Hours Day), reduced to 44 hours and then to 40. Paid public holidays became standard in the early 20th century and the number increased over time. Sick leave, long service leave and recreation leave came in. The last significant changes to affect the workforce as a whole were the extension of annual to four weeks in the early 1970s and the reduction of standard working hours to 38 hours in the early 1980s. Since then, with the exception of parental leave paid at the minimum wage, there’s been nothing. On the contrary, overtime, often unpaid, has become the norm. (My impression is that long working hours peaked in the 1990s, and have been scaled back as the labour market has improved, giving workers more bargaining power).
If a resurgence of labour saving technology is to benefit the community as a whole, and not just owners of capital, we need to resume the trend towards taking a large part of the benefits in the form of increased leisure.
I plan to write more about this, and the related ideas of Universal Basic or Guaranteed Minimum Income, in subsequent posts. In the meantime, go and read Tim Dunlop’s book.