Do new technologies, such as robots, destroy jobs and cause mass unemployment? Many current and past commentators have forcefully made this point in the public debate, but new research published in the Journal of the European Economic Association suggests that “technological mass unemployment” is indeed not something we should worry about.
Instead, our research documents the subtle shifts in the German labor market brought about by industrial robots, and highlights the crucial importance of functioning labor market institutions for coping with the challenges of technology-driven industrial change.
Starting in the 1990s, robots entered German manufacturing. Quite obviously, this has led to displacements of tasks. Many production steps previously carried out by humans, such as the assembly of cars, were now performed by robots. Yet, our key finding in the paper is that, firms did not fire those workers who were no longer needed. Instead, they retained and retrained (most of) them, and the majority actually ended up performing new, more complex, and ultimately better paid jobs.
In other words, the same worker who used to bolt cars was turned into a maintenance, software, marketing person. To be sure, firms could have fired the old car screwers and looked for younger workers to fill the new jobs. But German manufacturing firms decided differently, and went for “firm-specific human capital.”
Was that voluntarily? Sure, but with a twist. We find evidence that the ret(r)aining strategy was chosen more often in regions with higher density of union members. That probably reflects some sort of deal between management and works councils, who play an important role within the German system of industrial relations.
Ultimately, the deal was such that firms hung on to their incumbent workers, but in return unions promised to remain modest when it comes to future wage demands. This stabilized jobs for incumbents.
At the end, Germany has digested the rise of the robots much better than the United States, despite the fact that Germany (and also German manufacturing) is much more robotized than the US. In a related paper, Daron Acemoglu and Pasqual Restrepo find that robots caused severe job and wage losses in the American “hire&fire” labor market. Thus, in comparison, the corporatist German model seems to have performed pretty well when it comes to digesting this technological shock.
Further positive news is what happened to the young generation in Germany in response to the robots. Since firms mostly went for the ret(r)aining of old incumbents, it meant that fewer new jobs were created for young job market entrants in the manufacturing sector.
This meant that young entrants had to start their careers elsewhere. And they did, mostly in business-related services. We even find that young people, anticipating that fewer factory jobs will be available, responded to the robot shock by investing more in education.
The unlucky ones in Germany were those workers for whom the ret(r)aining strategy did not work out, possibly because they were employed in relatively weak and non-robotized firms, who missed out on productivity gains and lost market shares. But on average, we find no evidence that robots led to an overall decrease in employment or wages, i.e., on balance, the success stories in the labor market outweighed the unlucky ones.
Now, what may all of this mean for the future? Quite obviously, massive transformations are ahead in many industries, in manufacturing and beyond. No longer from robots, which are already a quiet, mature technology, but from other new digital technologies, such as artificial intelligence, and from forced technological change that originates in the necessity to reduce carbon emissions.
For example, what will happen to those workers who currently produce conventional combustion engines, or parts thereof? What will happen to small cities, whose local economy depends heavily on just a few firms that happen to be highly specialized in components for combustion engines?
Quite related to the key message of our research, a recent study by Fraunhofer claims that the overall number of jobs in the new, green car industry will be roughly the same as today. But it will be very different jobs—less bolting of turbochargers, more customer-services for self-driving electric vehicles.
On a positive note, our robot paper raises the hope that German manufacturing could also manage this transformation in a similar, non-disruptive manner. For this to work, on-the-job training and education will probably be crucial elements.
Feature image by Possessed Photography via Unsplash