MUNICH — Do machines replace humans? Since the beginning of industrialization 200 years ago, we earthlings have been plagued by this fear. From the early uprisings of the weavers to the 1970s "job killer computer" slogan, and up until the 2013 thesis of researchers Michael Osborne and Carl B. Frey, according to whom machines could soon take away every second job. But a German researcher, Terry Gregory of the Institute of Labor Economics (IZA), now presents a very different calculation. According to him, automation has brought Europe an additional 1.5 million jobs in the past decade.
Whether machines are our friends or enemies is one of the most intense economic debates, and one in which extreme positions dominate. On the one hand, you have the optimists, who only calculate models in which machines and people complement each other perfectly. On the other, skeptics the likes of Osborne and Frey who use vague job descriptions and neglect positive effects.
Gregory and his colleagues at the Leibniz Centre for European Economic Research (ZEW) try instead to get closer to reality using comprehensive models. Yes, between 1999 and 2010, machines cost Europe 1.6 million jobs, mostly of them in production. But according to what the companies had originally forecast, it should have been three times as much. On the other hand, computers and robots made it possible to produce goods more cheaply. As a result, consumers bought more and thus created new jobs. The companies made more profits, which in the wallets of their owners also turned into more consumption and thus more jobs. This resulted in the creation of a total of three million jobs — twice as many as the machines destroyed.
By 2025, digitization in Germany will create about as many jobs as it will destroy.
This positive balance can also be explained by the fact that factories are losing importance. Three out of four Germans now work in the service sector, delivering parcels, caring for the sick or designing houses. They are therefore not as easily replaced as people working in production on assembly lines, at least not yet. If, unlike Osborne and Frey, you look at concrete jobs instead of vague job descriptions, the automation potential of existing jobs drops from 50% to 10%, according to the OECD. Germany's Institute for Employment Research (IAB) doesn't see computers as job killers either. According to its forecast, by 2025, digitization in Germany will create about as many jobs as it will destroy.
This, however, doesn't mean in any way that those who lose their jobs to a robot will automatically get a new position that suits them. In certain sectors, the risks are concentrated. One in four Germans now works in a profession that can largely be handled by machines. Two million people work in warehousing, drive trucks or buses, which could potentially be replaced by autonomous vehicles. "Employees have to train themselves further, companies have to retrain them", Terry Gregory says. "Out with the routine, workers have to be redirected towards analytical, socially interactive activities. This is the transformation that economies must create." So, how much will the era of the machine affect people in the future? It will depend very much on politics. Also scientists at the World Bank have said so. According to them, the automation of the past decades has hit workers in the U.S., Great Britain, and Australia harder than elsewhere because of a bad social safety net and poor basic education that prevents laid-off industrial workers from being retrained for new jobs.
The consequences of the machine boom also depend on who owns them. Terry Gregory says that one positive effect is the fact that automation generates additional profits that the company owners spend, thus creating additional jobs. However, according to his findings, if a German company happens to be in the hands of foreign shareholders, they will spend most of their money abroad. The model shows that if all the profits had been drained away, there would have been 300,000 fewer jobs created in Europe. According to the model, it seems to make no difference whether Chinese, Indians or Arabs buy up most German companies, the end result is the same. There is, therefore, a strong argument for turning as many Germans as possible into shareholders, in order to share with them the profits brought by the machines, and thus stimulate national consumption.
Still, the bottom line is that, according to Gregory, automation in Europe has had a positive effect on jobs in the 2000s. The only question is: Will this also be the case in the future? This is the biggest counter-argument coming from skeptics like Osborne and Frey: The new technology of today and the future cannot be compared with anything we've seen before because the latter can potentially threaten every sector. Future robots, computers, and algorithms could simultaneously destroy simpler as well as more sophisticated service activities. They could deliver parcels as well as take care of the sick and answer customer inquiries in call centers. They could also design houses, prepare complaints and diagnose patients, thus replacing architects, lawyers, and doctors. That's already often the case in the U.S. If machines make products cheaper, the additional demand may not result in new jobs for people, simply because the machines would take up these jobs too.
Technologies always end up spreading more slowly than expected.
Terry Gregory finds this too pessimistic. "There have always been such horror scenarios over the past centuries," he says. Admittedly, artificial intelligence brings a lot of novelty with it. "But such technologies always end up spreading more slowly than expected." Until now, for example, only 5% of German companies have adopted networked robots and other such tools of the industry 4.0. "Of course, nobody knows exactly what the future holds. It's a gaze into the crystal ball. But people will have time to adapt to it."
They'll have to anyway. But according to a World Economic Forum study, not even half of Germans have the basic knowledge that the jobs of the future will require.
