The turbulent world of technological unemployment
- Only 10% of the 6.9 million jobs created by the US automotive industry between 1910 and 1950 involved manufacturing.
- Greater public spending on education could soften the downside of new technologies.
Spare a thought for the lowly iceman. In the late 19th century, natural ice was harvested from lakes, stored in ice houses and then transported to cities. Icemen would make daily deliveries to their customers’ iceboxes. At the turn of the century, Norway exported more than 1 million tonnes of ice each year. Ice-laden ships travelled to northern Europe, the Mediterranean, Constantinople, Africa – even as far as India. By 1920, with the introduction of artificial ice and mechanical refrigeration, Norwegian exports had fallen by 95% from a decade earlier.
Thousands of jobs disappeared, and ice warehouses were abandoned or converted to other uses. As an example of a job-destroying technology, refrigeration was just the tip of the iceberg compared to today’s robots and algorithms. “The feeling that technology is changing faster than ever before is repeated throughout the centuries,” says Louise Skyggebjerg, an expert in the history of technology at the Technical University of Denmark. “In the 19th century, we saw this with the proliferation of railways, steam engines, telegraphs and new factory machinery.” To some, worries about ‘technological unemployment’ may seem odd – after all, technology has created millions of new jobs worldwide.
As the IMF noted in a recent report, “Technology boosts productivity, which in turn drives strong per-capita GDP growth and has been associated with expanding employment.” But, the report continued, “Gains in employment and income can come in spurts and tend to favour different sectors over time”. Set against this context, understanding how the two biggest inventions of the 20th century shaped work may help predict the impact of automation in this century.
Automobiles: from the factory floor to family holidays
Thanks to Henry Ford’s introduction of the assembly line, productivity rose quickly as the auto industry developed. In 1909, 10,666 Model Ts were produced; by 1923, this had risen to an astonishing 2,011,125. During that period, car prices went down, demand increased, and so did worker wages. Ford’s methods were not only copied by his competitors, they quickly spread into other manufacturing industries. According to estimates by McKinsey Global Institute, the car’s popularity created 6.9 million net new jobs in the US between 1910 and 1950. Yet only 10% of these were in manufacturing. Dealerships, repair shops, petrol stations and convenience stores added 30%, and transportation and logistics a further 25%.
Then there was tourism. “The rise of the car created tourism in areas that were previously difficult to reach – ‘windshield wildernesses’ where roads were built to show off the natural landscape,” Skyggebjerg notes. The new family-holiday concept created demand for new roadside attractions: motels, campgrounds, drive-in movies and restaurants, shopping malls and parking lots.
Computers: no typing pools, but increased stress
It is rarely easy to predict future demand for a technology. In 1943, IBM president Thomas Watson famously remarked, “I think there is a world market for maybe five computers.” A few decades later, computers and word-processing software had done away with typing pools and virtually eliminated the profession once called ‘shorthand-typists’. “Human computers (used for calculation), researchers, state officers, statistics officers, industry payroll, accounting, administration, bankers and insurers were all affected – really any data-processing role came under risk,” says Karin Zachmann, a technology historian at the Technical University of Munich.
Even so, early computers created new work. “They led to many new jobs mapping and translating real-world processes into digitised language,” Zachmann notes. “And we should not forget applications that were entirely new: computer-games engineering, for example, became a new specialisation in informatics.” Disruptive technologies don’t just save on labour, they fundamentally change norms within industries. “We no longer talk about secretaries, we call them administrators,” Zachmann says. “It’s a very different from the secretaries of the 1950s and 60s. On one hand, the tasks are more complex, but on the other, there is increased stress because the changes in communication from letters to emails increased the amount of correspondence and the speed at which people are expected to respond.”
Education and redistribution
Low-skill mechanical jobs disappeared from all aspects of life throughout the 20th century: lift operators, train brakemen and switchboard operators were all replaced by circuit boards. The ongoing maturation of machine-learning and other forms of artificial intelligence now means that automation is creeping into mid-skilled ‘cognitive’ tasks and even some highly-skilled areas. But Skyggebjerg sees no reason to panic. “We have yet to see the many predictions of mass technological unemployment come true,” she notes. “But on an individual level, of course, many people have already experienced unemployment due to new technologies”.
The future of work may be unpredictable, but governments can still immunise their workforces from the biggest risks. In its report, Technology and the Future of Work, the IMF ultimately concludes that two policies could soften the negative impact of change in developed economies: (1) increasing public spending on education to reduce the number of low-skilled workers, and (2) a redistribution through income-tax cuts targeted at middle-income workers. “Both policies maintain gains from the underlying advances in technology, while distributing them more evenly,” the report concludes.