In the late 1990s, when Anna and I were living in France, I got into a very wide-ranging argument with a couple of late-20s but astonishingly right-wing Americans who were staying in our Parisian hotel. One of the things we argued about was unemployment; they were utterly convinced that jobs existed for everyone, if they wanted them. They pooh-poohed my suggestion that technology had already eliminated or severely reduced the number of unskilled and low-skilled jobs (e.g., potato-picking, hay-baling and many other agricultural tasks) and was displacing even moderately-skilled workers (e.g., ATMs instead of bank tellers) simply because machinery costs less than labour.
History shows that technological change has powerfully affected employment. In 1500 agriculture employed an estimated 75% of the British labour force, but by 1800 that figure was 35%. When manufacturing began during the 18th century it was mostly in small-scale operations within the home or small workshops, but by the end of the 19th century people were crammed into huge factories in large industrial cities. Automation and steam power made these transitions possible.
It’s certainly true that economists have long held that raising productivity means that any automation which economises on labour will increase incomes, thereby generating demand for new products and services and creating new jobs for displaced workers. This is a long-term view, and isn’t much comfort to the worker replaced by a machine (remember how many branch staff banks used to employ?) Nevertheless, thanks to industrialisation, average British incomes tripled from 1570 to 1875 and more than tripled from 1875 to 1975, creating employment that more or less matched the rapidly expanding population.
Our current era of increasing computer-driven automation could have rather different results. Economic inequality has risen steadily in most Western countries since the 1980s. Owners of capital reduce investment in workers through automation, allowing them to capture ever more of the world’s income, while the share going to labour has fallen. In rich countries the typical worker’s wages, adjusted for cost of living, are stagnant. In the USA the real wage has hardly changed in forty years, and in Britain and Germany, despite high employment, wages have not risen in a decade. Because people in full-time employment are working longer and longer hours, the stagnant average wage conceals a growing underworked class who earn less and less.
According to some researchers, nearly half of job categories will be open to automation within two decades. Of course, some jobs – especially those currently requiring high levels of education and delivering high wages –will survive. Tyler Cowen writes in his most recent book, “Average is Over”, that rich economies are bifurcating into a small group of workers with skills highly complementary with machine intelligence (for whom he has high hopes) and the rest.
Everyone should be able to benefit from productivity gains achieved through technology. However, society will be sorely tested if growth and innovation continue to deliver substantial gains to the highly skilled while employment opportunities at stagnant wages dwindle for everyone else.
*This post includes edited material originally from The Economist, 18th January 2014.