 Welcome to the European Central Bank podcast bringing you insights into the world of economics and central banking. My name is Michael Steen. The digital revolution is often associated with labour markets and seen as potentially destroying jobs. In the last episode we did some myth-busting on this and explained how technology is so far not that big evil robot that comes to replace our jobs. On the contrary, digitally intensive sectors can actually make important contributions to employment growth. We also explored differences between European countries and their digital economies. Today we want to dig a little deeper on the impact on jobs. We want to better understand how technological progress affects the way we work and explore the longer term implications of digitalisation on labour markets. Our first guest today is ECB economist Robert Anderton who joined us in the last podcast episode as well. So Robert, thanks a lot for coming back and welcome to the podcast. Thanks for inviting me. In the last episode we spoke about how digitalisation can lead to more jobs in the medium term, contrary to the common belief that technological progress can destroy jobs. And that's one of the results of the paper that you recently published. Now, this concept of job polarisation is one of the effects of the digital revolution on labour markets. And I was wondering if you could start by explaining what is meant by job polarisation. So when economists talk about polarisation in the labour market, we generally mean the ongoing declines in the employment share of middle-skilled jobs compared to high and low-skilled jobs. This process is associated with digitalisation. What was seen over past years is the automation of routine tasks. Jobs characterised by a high content of routine and repetitive tasks can typically be performed more efficiently by machines or computers. This has resulted in a fall in the share of middle-skilled jobs such as bank tellers, machine operators and office clerks. Meanwhile, the share of non-routine tasks is rising. This corresponds to an increase in the share of high-skilled jobs. These are particularly jobs that require complex analytical skills with limited automation potential, which tend to be complementary to digital technologies. On the other hand, automation has not yet affected many non-routine, low-skilled jobs, those that typically do not require specialist education but do require things like intuition, flexibility, adaptability or interpersonal interaction, which are hard to automate. Think of cleaners or security staff. So overall, job polarisation is where automation of routine activities has led to growth in the share of low and high-skilled jobs at the expense of middle-skilled jobs. We talked in the last episode about differences in European countries in terms of the digital economy and how digitalised they are. Some are highly digitalised, some are lagging behind. So looking at job polarisation, what are your observations here? The work on polarisation in our paper shows that over the past quarter of a century, the share of non-routine tasks have been rising, while routine tasks have been falling in Europe. This is consistent with job polarisation and may be associated with increasing digitalisation and automation. Job polarisation is alive and kicking in Europe. We also investigated whether the decline in average hours worked in some EU countries over recent decades has exacerbated or mitigated job polarisation. We found that hours worked have fallen more for some routine jobs compared to non-routine jobs and that the decline in hours worked is exacerbating the impact of job polarisation at the top and middle parts of the skill distribution. Looking forward, various studies estimate that over the next few years a significant proportion of jobs could be automated, while substantially more could undergo significant change as a result of automation. Estimates of how many jobs are at risk differ greatly from around 10% to 60%. There are also big differences across countries in terms of how many jobs are at risk of automation. Looking at some studies, it sometimes seems to me that the countries who may be most affected are those who are currently the least digitalised as the more highly digital countries have already at least partly gone through this process. But of course, as the previous podcast on digitalisation made clear the automation of some jobs will most likely eventually be at least offset by the creation of new jobs. There's also the role of artificial intelligence and people have seen recently in the news there are some now incredible AIs that can even write texts that it's hard to distinguish them from being ones that are actually written by humans. So can you talk us about artificial intelligence and what we think that might mean for the future of jobs? Well, the future may not be the same as the past as increasing use of artificial intelligence may also lead to routinisation of high-skilled jobs. AI algorithms particularly excel in prediction so they can replace humans in prediction tasks which are normally thought of as high-skilled jobs. But of course, you still need humans to decide how to use the predictions although even automating decisions has now become increasingly possible with AI. The legal profession is a perfect example. AI and other technologies can now do many of the tasks that we previously thought could only be done by highly skilled and qualified lawyers. AI software can be used to forecast legal outcomes for instance by finding and reviewing past relevant case law very quickly taking account of past decisions of an individual judge and predicting the likely result of a case. Similarly in healthcare and medicine AI is increasingly used for automating medical diagnosis. Overall, AI can allow the automation of an ever-increasing range of tasks turn in many high-skilled jobs into routine tasks. We don't yet know if this could possibly upend the relationship where technology has historically had a net positive effect on job creation. Okay, thanks a lot for your insights there Robert. Now we are going to turn to job opportunities. So one key aspect in the discussion of how the digital revolution affects labour markets of evolving tasks and more job growth is the so-called digital platform economy. And my next guest is Lara Vivian, also an ECB economist and she's looked into this topic in some detail. Lara, welcome to the podcast. Thank you for having me here. Okay, Lara, so you've taken a closer look at the platform economy. Help me out here, what does this term actually mean? The first challenge of the platform economy actually lies in its definition. So what all platforms have in common usually is that the two sides of the market match through a platform, an online platform. So the broad term here is sharing economy or platform economy. And they can include the platforms where, for instance, you can look for a place to rent, a chance to invest in a new product, also a specific task such as the development of a software or a ride back home. So if a task is involved such as the development of a software or a ride back home, then we say that the paid labour that is necessary to complete this task is platform work. And then there's this other phrase that I have heard of. This is the gig economy and that's linked to that part of it. Is that correct? Exactly. So the term gig economy tends to be a bit more blurred. However, it gathers the platforms which rely on platform work. Okay. And so, and just to be very super concrete here. So platform economy, that would include just use some household names. Obviously, the ECB doesn't take a stance on which ones are good or bad or anything, but so Airbnb, that's platform economy. Uber, getting one of those scooters that they're left lying around on the roadside. That's all part of the platform economy. So we know what we're talking of. Precisely. That's correct. Now, how has the platform economy evolved recently as societies are becoming more and more digitalized? Platforms tend to rely on algorithms for their business models, for instance, but also the increasing capabilities of the Internet, the decreasing prices of digital devices such as smartphone and tablets. Of course, they've all contributed to the mass adoption of those devices and in turn to the growth of the platform economy that we have observed over the last decade or so. Okay. Because a lot of this is happening on a smartphone very often, isn't it? So the size of the platform economy has increased rapidly, but how do different countries in Europe compare and how do we compare against other major economies? So in the occasional paper, we review some estimates on the platform economy for Europe and according to one study, in 2016, platforms accounted for 0% to 2% of GDP in Europe, which is obviously quite small. Belgium was the country in which platforms only accounted for 0.104% of GDP and while the largest share is about 1% in Estonia. Overall, we do seem to see that sharing economy are larger in Eastern European countries compared to Southern European countries. And in Estonia, as we heard in the last episode, is also a very highly digitalized economy anyway. Although largely driven by the sort of e-government and stuff there, right? Exactly. Okay. So let's go into maybe a bit more detail about what a typical job looks like in the platform economy and some of the possible policy challenges that this brings about. So I want to talk about the survey conducted by the Joint Research Centre of the European Commission in 16 European countries. And they do find that in 2018, around 1.4% of the workers engaged in platform work as the main job. And this is excluding people for whom platform work might be a secondary or seldom activity. So of course here as well, there is some heterogeneity. So the average worker tends to be male, young, educated and works few hours. So they're not doing this full time then? Exactly. If we include in the share, people, workers that might be engaged in platform work also seldomly, then the share goes up to even 18% in countries like Spain. However, depending on the platform they work in, they might be very satisfied with job flexibility and we see this for instance in platforms where they perform online tasks. Although depending on the survey, a significant percentage of workers may report that they would like to work longer hours. The degree of autonomy that they benefit from also depends on the platform. In some platforms, workers can choose what tasks they want to carry out. In other platforms, it is going to be assigned to the worker by the algorithms. So often among the issues which are raised is the worker status and therefore the level of social protection and contribution, which is often not included in the compensation of the platform worker. There are also additional issues regarding the rules on collective bargaining and access to training for instance. So what are some of the policy challenges that this might bring with it? So similarly, as with most tech companies, there is a risk of the possible emergence of dominant platforms and policy questions also arise on how to promote fair competition without offering the innovation of the platforms might bring. Some policy tools are already in place that could be used such as the regulations of mergers and acquisitions. However, some challenges are specific to the platform economy and there are new areas of policies that might need to be tackled such as, for instance, the role of algorithms in the business models and their transparency, the rating system, or there might be a need to regulate whether workers can and how they can switch from one platform to another, for instance. I will mention here that policy institutions such as the International Labour Organization or the European Commission have launched agendas precisely to tackle the policy challenges which are specific to the platform economy. Okay, so it's still very much a developing area and obviously something that needs to be looked at quite closely. Lara, thank you very much. Thank you. Okay, now we'd like to turn back to Robert for another question which is related to the coronavirus pandemic and the lockdowns which of course have seen a massive shift in the way we work towards much more flexible working arrangements with people suddenly working from home where they previously worked in offices. Just at the ECB itself we saw that more or less from one day to the next about 4,000 of us started working remotely. What are your observations on this? Could this be something that gives us a necessary push towards faster digitalisation? This is a really interesting issue right now. It's a bit speculative but it is possible that the COVID-19 pandemic could result in a permanent acceleration in the take-up of digitalisation. Sometimes this is on a scale and at a speed we would not have thought possible a few months ago and which is unlikely to be fully reversed. We've seen rapid increases in online sales of food, consumer durables etc. which may have implications for offline sales. Growth in video conferencing and teleworking may reduce international and even local travel. This can have implications for the mix of jobs on offer and could lead to some temporary disruption in the labour market particularly if some workers don't have the necessary skills for the new jobs. On the other hand though, the massive increase in the use of teleworking now makes it easier for virtual labour flows to move within and across countries to wherever labour is in demand potentially reducing unemployment. A more rapid take-up of digitalisation can also increase productivity making the economy more dynamic and increase growth and boost employment. Undoubtedly, some of the rapid increases in digitalisation in response to COVID-19 are unlikely to be completely reversed. This may provide some challenges for EU countries but it also potentially offers many opportunities for EU countries to catch up with major competitors in the adoption of digital technologies. In 2015 the European Commission announced a digital single market strategy as part of the Digital Agenda for Europe 2020 programme. Some say that the European digital single market has the potential to become one of the most valuable trade markets in the world for online business. What are our thoughts on that? The digital single market is clearly a very important policy strategy. No progress has been made but its implementation could be possibly accelerated and expanded. Let's not forget that at the EU level the aim is to further the single market in the digital domain and to facilitate the process of digitalisation across the EU. This would really help to deliver increased choice and lower prices for both consumers and firms as a result of scale economies and improved EU competitiveness. But there is also potential for more investment by the public sector in AI and robotics and faster progress in EU government. Similarly, digital startups and investment in intangible capital could be stimulated by greater provision of venture capital. Not all countries in the EU are near the technology frontier in terms of the dissemination and adoption of digital technologies. And firms at the very forefront of the digital revolution are often found outside of the EU. Perhaps a more far-reaching digital policy agenda is needed to overcome second mover disadvantages, advance closer to the technology frontier and to remain competitive in global markets. Thanks very much for those thoughts. Thank you. And thanks also to Lara. This brings us to the end of today's episode. We've discussed how digitalisation affects our tasks and learned about job polarisation. We touched on the role of artificial intelligence, talked about the challenges ahead with the growing so-called platform economy, and we explored how the current coronavirus pandemic can change the way we work for the better. This was the second episode of our podcast series on the topic of digitalisation, so do check out the first one if you missed it. And stay tuned for more on the topic later. We'd love to hear your feedback and thoughts for future episodes via social media. You've been listening to the ECB podcast with Michael Steen. If you like what you've heard, please subscribe and leave us a review. Until next time, thanks for listening.