Robotisation reduces the wages and hours worked of people who can easily be replaced by robots, while increasing the wages of those who control the machines, according to new research by Professor of Urban Economics Hans Koster. Together with Daron Acemoglu (Professor of Macroeconomics at MIT) and Ceren Ozgen (Associate Professor of Applied Economics at the University of Birmingham), he studied data from Dutch companies and workers.
We interviewed Hans Koster about this new research. ‘Industrial robots are autonomous machines that are used in manufacturing in order to produce goods’, says Koster. ‘These robots replace specific tasks normally performed by humans. You would expect that this is a good thing. It’s a technological improvement. It makes the labour market more efficient, leaving people with more time and money for things that are fun. But are these effects the same for all groups of employees? We expected that the effects for people who perform tasks that are taken over by robots will be affected negatively, and that other people will benefit because those robots are complementary to activities they already perform.’
Replaceable tasks – decreasing wages, reduction in working hours
And that expectation has come true. ‘We used Dutch microdata about employees who were tracked over time. We looked at their wage development over time, their chances of working, their chances of becoming unemployed. We can distinguish between employees who perform tasks related to robots and people who perform tasks that are directly replaced by robots. And then we do indeed find that people who perform replaceable tasks are faced with a relative decrease in wages and slightly fewer working hours. People who are not directly replaced by robots benefit from it. They see their wages rise, work slightly more hours, having slightly more employment.’
An example? ‘You can think of IT-like jobs, people who have more work because they have to keep an eye on those robots. But it might also be the case that the robots make processes faster, creating more work on a later stage in the production process, for example in administration.’
Labour market inequality
Why is this research so important? Koster: ‘Our co-author Daron Acemoglu has been studying the diverse impact of automation for years. Economists tend to think that all automation is great. But it is not that simple, there is more to it. There are clear disadvantages, and often for a group on the labour market that is already struggling. The people who perform replaceable tasks are usually the low-skilled and low-paid workers. As a result, automation increases inequality in the labour market. This applies to industrial robots, it may, however, be different for other sectors.’