Daniel Nippard, Project Director of the MarketMakers project in Bosnia & Herzegovina, interviews Stijn Broecke of the OECD's Employment Division about the future of work.
Daniel: We hear a lot about automation and its impending impact on labor demand. Are the risks of widespread displacement of people by machines grounded in evidence? Will this hit Generation Z in a big way in the course of their working lives?
Stijn: It depends on which sources we refer to. I think that some estimates that say that 50% of jobs will disappear are probably exaggerated. On the other hand, I don't think it is exaggerated to say that a significant share of tasks will be automated. The real question is: how fast this happens, who is impacted by it, and how do we help these people get into new jobs? What we have seen in the past and what we might see in the future is that this also creates new jobs. We shouldn't panic though. Yes, more jobs will be automated, but you'll get new ones appearing.
Daniel: Yes, automation can be opportunity-generating as well. The pessimistic view assumes a net reduction in employment from automation. But, do you think the two are somewhat symbiotic – is there a view that growing automation could produce a net increase in employment?
Stijn: I don't have a crystal ball. One of the best ways of looking at the future is to look at the past and even what is happening in the present. When you look at employment rates in most OECD countries, they've been rising and are at record highs, despite automation. There are no real signs of net job destruction. Policy-makers from many of our member countries are rather more concerned about shortages of people and skills. So, my sense here is that we'll have higher job creation and for a number of reasons: one, we will create entirely new occupations that didn't exist; two, in the future, we will be making goods cheaper and better, people will be more productive, wages should rise – leading to increased consumption and likely more work in the end. At least in the foreseeable future, we don't have to worry about 'technological unemployment'. With one caveat perhaps – this might be true globally, but you might see that in some regions where you have a concentration in a certain type of industry or occupation prone to automation, massive unemployment is the likely result.
The key challenge will be to manage transitions. You'll get new jobs being created but those are not the same as those that are being destroyed – they're in different occupations, different sectors, and often even in different geographical areas. So, at a local level, you may well see a reduction in jobs and the significant hardship that goes with that.
Daniel: You've predicted a little bit my next question – on the regional and localized effects of these changes. I appreciate that the Western Balkans region is perhaps not your number one focus at the OECD, but here on the southern edge of Europe, is it fair to say that we might be hit a little harder?
Stijn: I don't have strong evidence to hand here, but I can speculate. I suspect there are more of the kinds of jobs that can be automated, but I guess there is also a brake on that. As wage costs are lower in many of these countries, and as long as that is true, then it may not make business sense to replace a worker with a machine. There are risks here, though. When it does happen, the social costs will be higher because you're talking about greater numbers of people, and the social protection systems and training systems are not as well-developed as in other countries. The second thing is that when it happens, it may already be too late as these countries will have fallen further behind the countries that automate sooner. So, the challenge might be that automation is not happening fast enough, rather than that it's happening too fast.
Daniel: I have read that these days it is often the semi-skilled work, as opposed to the skilled and unskilled work, that is among the first to be automated. In countries with above-average growth in automation, unskilled and skilled employment is not only stable but increasing.
Stijn: Most of the tasks that we can automate now are at the lower-end and in the median of the skills distribution so that is where you will have the greatest impact. We've seen polarisation already, though going forward I suspect that automation is even more likely to happen at the bottom end of the skills distribution – so whether this polarisation will continue in the future, I'm not sure.
Daniel: Maybe those semi-skilled positions in the middle of the skills spectrum have been the first to go because of the point you made earlier on wage costs – there was less of a need to experiment with automating unskilled, cheaper tasks?
Stijn: I think that's very possible. I guess also one issue about jobs at the bottom of the skills distribution is that some of them are very difficult to automate. If you think about the service sector, there are human interactions and often human decisions that still need to be taken. Some of these are difficult to automate and will be for some time I suspect.
Daniel: Yes, it's not necessarily a straightforward correlation between the skill demands and complexity of a job and whether technology is ultimately used to replace that job.
Stijn: Right, like if we think of a waiter, which is supposedly a low-skilled job – actually it involves lots of skills that machines can't do. Similarly, for nursing and jobs like that, there is a 'human touch' that is very difficult to automate. Perhaps our notion of what constitutes a high- and a low-skilled job needs to be revised.
Daniel: So how can Generation Z future-proof itself so as to not be obsolete before the end of their working lives? What skill sets are required among labor market entrants today in order to get all the way through to retirement age without having too much time outside of work?
Stijn: I'd be rich if I knew the answers to those questions! There are perhaps a couple of points that I would like to make. In terms of the types of jobs that we might see in the future, we know that some sectors are likely to grow over the coming decades. With an aging population, you have a growing need for health and social services. There is also a move towards leisure services as people in retirement have money to consume more. We know that social skills and soft skills will become more and more valuable. Therefore, in terms of skills demand, young people can to some extent situate themselves. It doesn't mean that they will be more valuable than having good engineering or mathematical skills, but we predict they will grow in importance and help you to keep a job.
We also see that young people in the labor market are increasingly struggling with a number of issues – one of these is under-employment. Globally, we don't necessarily see an increase in unemployment, but we do see an increase in under-employment and this affects young people more than other age groups. At the OECD, we have also been looking at mobility in the labor market, and how long people stay in a given job. We find that people are staying less and less long in their job and this is particularly true for young people. We're not entirely sure if this is their choice or not – I suspect it's a mixture and that there are some external forces that are pushing people to move more often than they did before. Finally, automation tends to affect entry-level jobs more than others. We find that young people, especially those that are low-skilled, find it more and more difficult to get well-paid jobs and are being pushed down the pay distribution.
Daniel: I'll come back to one of the first points you made in your answer on the types of professions and skills demands of the future. I believe it is increasingly common for people to think that the next generation will all need to be software engineers in order to prosper in the labor market of tomorrow. But, common sense would suggest that there will continue to be a great diversity of professions and therefore skill requirements.
Stijn: Yes, I suspect that there are jobs like hairdressers, gardeners, plumbers, electricians, for example, that will always be there – I don't think jobs like that will disappear overnight. It also comes back to a choice that we need to make as a society. It's not because the technology is there that it necessarily will be adopted. People may always prefer having a real person as a waiter rather than having a robot. So, if we don't like it as human beings, then we won't choose it. There was an interesting survey done by the European Union into how comfortable people were with automation. It found that people were very comfortable with automation when it comes to space exploration, but they don't feel comfortable with automation when it comes to caring for the elderly or for young children. Someone once said that we shouldn't think about automation as what is possible or not, rather we should think about what kind of jobs we don't want to be automated and that's much more relevant.
Daniel: Thank you for joining us Stijn! We've covered a lot and I've learned a lot!
Readers can access the OECD's latest research findings on the topics discussed above and far more in the Employment Outlook 2019.