Ryan Avent
Ryan Avent, thank you very much for joining us today. You’ve got a new book out entitled ‘The Wealth of Humans: Work, Power, and Status in the Twenty-first Century’, and it addresses one of the hottest topics around at the moment – the digital revolution, and what it means for the future of work.
So, what do you think? What does it mean?
Well, I think it means big change. I think the starting point of the book is that the digital revolution actually is probably going to be as transformative as the industrial revolution and the big technologies like electricity and steam that we saw then were. I think this transformation has already begun, and ironically, the evidence of that is in the struggles that we’re seeing across lots of countries that workers are facing in terms of limited growth in wages, in terms of rising inequality. These things, as the book argues, are all connected to the growing role of these powerful new digital technologies, and especially of machine intelligence.
How many jobs do you think will actually disappear? Because if you look around in the economics literature, and the policy literature, you see a lot of conflicting studies. Some studies claim that about half of the jobs are susceptible to automation; others say it’s 10%. I personally would argue it always depends on the assumptions, and we actually don’t know. But what is your view? How many jobs are likely to be affected?
Well, I think you’ve got it right in that it really depends on what your assumptions are. If you just look at the task content of jobs, and you assume that jobs are not going to change in the balance of tasks that they consist of, then I think probably a lot are quite vulnerable. There’s the one paper from scholars at Oxford that says 47% are potentially automatable over the next couple of decades. But the thing is, jobs change over time, so as a worker has some of the functions that they do during the day, some of the things they do on a computer or talking to customers, as some of those things are automated, what their responsibilities are end up shifting. They may find themselves doing a more particular kind of work.
So, I think in practice we won’t see as much automation away as some of the more dire forecasts warn. What my book tries to point out though is that in fact the biggest effect is not going to be mass unemployment. The biggest effect of the digital revolution is not going to be massive numbers of workers who just can’t find any work; it’ll be that the work they find ends up being very low-paying, because the displacement effect of these new technologies is so great, and the economy is asked to absorb so many new workers, that that’s just going to put an incredible amount of downward pressure on wages. That’s the real short-run challenge, I think.
Right, and other than that, it’s the mixture between substitution, augmentation, and creation, isn’t it? So which jobs will disappear altogether, which jobs will have a completely different skill profile which might also have an impact on the quantity of people needed – think checkout tills at supermarkets – and obviously there will also be new jobs… I mean, the job title ‘Social Media Manager’ is a rather recent phenomenon.
But the question is also that you mentioned about inequality, it’s the one about social mobility. The big question here seems to be, and this links back to a discussion we had with David Autor from MIT, just assume, say, the single job in terms of numbers that most people do in the United States seems to be truck driver, and if you just think about self-driving trucks, what would happen to all these people? They’re unlikely to move up the value chain and become software engineers, and there’s a limit to what kind of retraining you can do to how many people in what kind of timeframe.
So, what do you think all this might mean for social mobility?
I think it’s kind of a grim outlook. We have a sense of what it might look like from the decline in a lot of manufacturing industries over the last 30 years or so. What we’ve seen is that there just hasn’t been nearly as much movement from declining places to growing places, from declining industries to growing industries, as we would have expected. And so, as you start seeing entirely new kinds of job categories being affected, and truck driving, taxi driving, these are things that employ millions of workers across America and Europe, the challenge there is going to become much more significant.
I think we have to be first of all more proactive in our labour market policies. In America particularly, but also to some extent in some European countries, there was the attitude that if you had labour markets that were flexible, a lot of this stuff would take care of itself, and I think we totally have to rethink that.
But I think one of the flashpoints over the next few decades is going to be the fact that good jobs for a lot of workers without high levels of skills are going to be available in cities and in countries that are doing well. So if you are a truck driver who has lost your job, you might not be able to retrain as a computer engineer, but you might be able to get a job as a plumber or an electrician. That job’s only going to pay a lot if you’re in a city like London, or Frankfurt, that is otherwise thriving. The problem is, people in London and Frankfurt don’t want to build enough housing to accommodate everyone, they don’t necessarily want to have people from other countries coming to work in close proximity to them.
And so moving people to these places is going to be hugely important for making sure their standard of living keeps growing, but also very difficult politically and socially.
So at the same time, when such movement might become an economic necessity, what we might want to call a cultural backlash against globalisation might prevent just that.
Absolutely. I think that’s one of the big risks… We’re already seeing this play out, aren’t we, within the European context. This is certainly the subtext of the Brexit debate, I think.
Yes, absolutely. There’s clear evidence that there’s a bigger cultural problem against change underway, because it’s now almost completely divorced from underlying factual circumstances. That no longer is one of the key issues, so you also have to address the emotional and feeling level where people just feel disaffected and are rebelling against what is going on.
But what I quite like about your book as well is that you started with a personal anecdote. Obviously journalism and writing is an industry that is also affected. Maybe as a side note, what do you see going on in your industry and where do you see it headed?
Well, I think it’s a nice sort of microcosm of what’s happening more broadly, in that we’ve… There has been a lot of disruptive change that has made old and established companies uncomfortable. As the industry has adjusted to that, we’ve seen some companies make an awful lot of money, and others go out of business, and that’s an important thing to keep an eye on.
At the same time, you have different effects depending on what kind of workers you’re looking at, what class of skilled labour you’re looking at. And so I think for the people who are really doing amazing storytelling, who have scarce skills in terms of gathering facts and creating something that people want to read, they’re going to be okay for at least the next 30, 40 years. It’ll be a long time before computers can really encroach on that. But the people who are writing beat stories, financial reports that require more modest level of skills, those jobs are very vulnerable. They’re vulnerable because automation is increasingly a threat. They’re vulnerable because they’re kind of the most marginal sorts of work, and so companies that are under financial pressure are going to start trimming those jobs first.
It’s interesting to see what’s happening to the industry as a whole, but the distributional effects are something that are particularly worrisome, especially for people trying to get a foot into the door, who are just starting out in the industry, young people. You have to already have a lot of personal wealth in your bank account to be able to really have a go at the industry, I think.
What I really liked about the argument you constructed is that you tried to tackle one of the big mysteries head-on. We see a lot of writing and policy discussion about a digital revolution and how it’s going to change lives and our economies, but at the same time if you look at productivity figures, they’re stubbornly stuck, and in some areas actually declining. You mentioned that the key point that you would try to achieve is reconciling this kind of technological change with rising global unemployment, and the disappointing growth and productivity figures. So what is your key argument, how do you actually reconcile this?
This is the fundamental question, isn’t it? If technology is so great, and it is accelerating as we think it is, then why is productivity so dismal? I think the key is in looking at what’s happening to labour. What I say is happening is that technology’s creating an abundance of labour; it’s a labour glut. That’s occurring in part because we’re increasingly able to automate lots of different kinds of work; it’s occurring in part because technology has allowed countries like China and India to participate more fully in the global economy, and that’s brought in hundreds of millions of new workers. And it’s also happening because in some industries, and journalism is one of them, the return to being a highly skilled worker is increasing, and those workers are able to capture a lot of value that might normally have been spread over many more workers.
And so the upshot of all this is that there’s just a lot of effective labour available to firms, and that reduces labour bargaining power, it places a lot of downward pressure on wages as the way that we find jobs for all these willing workers. And ironically, if you’re an employer and you’ve got as many workers as you want at a low rate of pay, you don’t feel very much pressure to go out and invest in new labour-saving technologies. So if you are managing a warehouse, you have lots of people willing to work at a low wage, you don’t need to look into hiring robots to automate all the jobs stocking the shelves.
If you are operating a shop and you have cheap cashiers available, you don’t need to automate the tills. So the funny thing that we’ve seen – it’s not actually that funny, it’s upsetting – is that low productivity growth is resulting because employers don’t have to invest in these productivity-enhancing technologies because there are so many cheap workers around.
Up to the point where the technology price falls that much that it’s actually becoming even cheaper than the cheapest labour. If you look at instances like, I think it was Foxconn in China now replacing about 20,000 already lowly-paid workers with robots because they work 24/7, and basically repay their cost within a year. No human being would be able to compete with that.
Right. And the robots aren’t going to go on strike, there won’t be a story about the horrible conditions that robots are facing. So I think you’re right that there’s a bit of a race to the bottom that humans are bound to lose, and as this occurs the pressure on labour markets will intensify, and the demands the policymakers face to try to do something about it will grow.
Well, we’ll get to the potential solutions in a minute. But maybe one of the rarely addressed sides of the digital revolution is that it might actually do more to how people live than to their economic lives. One of the arguments, I can’t remember who said it, but somebody said, “We wanted flying cars, and we got 140 characters on Twitter.”
I’m sometimes wondering whether there is a connection between the cultural backlash against globalisation and the way how these technologies just change how people live on a daily basis, especially the flow of information, so that makes people retreat somehow. What do you think is the cultural effect of all this, beyond the economics?
Well, it’s hard to measure, but I think it’s significant. If you think back to the industrial revolution, or if you imagine what it was like then, there was this kind of palpable sense that things were changing rapidly, the world was being turned on its head. You had huge improvements in the speed of personal mobility technologies. You had skyscrapers growing out of nowhere. You had electric light. All these things kind of led to the sense that the world is totally different, and alongside those big changes you had huge cultural change, and you also had a huge cultural backlash against that change, and there was a lot of social unrest at the time.
I think we are coming out of a period where there wasn’t a lot of that sort of dramatic technological change, and emerging into one where there was, and we see it in terms of how people are able to follow events around the world on their phone; the phones themselves, which kind of upend all sorts of personal interaction; and then also the way that things are changing in the workplace and the stresses that they’re placing on jobs.
I think we’re only beginning to realise how fundamental our daily lives are changing, and will continue to change. Whenever you’ve got that kind of a shift, there has to be a process through which you come up with new norms, new ideas about what’s acceptable. Is it okay to treat someone online on Twitter with… To hurl nasty expletives at them the way that some people do, or do you need to come up with new norms about online behaviour?
I think there’s the vague sense that things are different and we all kind of feel a little uncomfortable, and that’s a very industrial revolution sort of feeling, and it’s only going to get more intense, I think.
Yes, we are certainly in a transitional phase, and most likely only at the beginning of this. Obviously one dimension that is not very often talked about is the velocity of change rather than the scope of change. If this all accelerates even further, that will only add to the insecurity felt by many people.
Yes. You mentioned driverless trucks – just to start imagining how our daily lives would change if a significant share of the vehicles on the road were driven by computers. The amount of change in terms of norms, in terms of what we expect from personal responsibility, what’s seen as acceptable behaviour. Is it okay to just allow the driverless car to drop your child off at football practice?
I don’t know, it’s mind boggling, and it will be exciting to watch unfold, but there will also be a lot of moments that are really uncomfortable.
People will be very anxious to follow all of this. But let’s move on to the very “easy” part of the conversation: what can we do about it?
We mentioned the cultural dimension, and especially also the economic dimension, what it means for the world of work; what it means macro-economically is also a major issue, because the digital revolution is above all a supply-side revolution, and it might add to existing demand-side problems. So what do you think the key policies are that should be pursued to manage and shape this process?
It’s an incredibly complicated sort of thing to work through, because a lot of the things that seem like good policy ideas, along one dimension can make things worse along others. But I think the fundamental problem that we’re seeing right now is the fact that purchasing power, which we’ve always allocated to people through work, for the most part – you earn your purchasing power by going out and getting a job – that’s no longer functioning the way we need it to, for a lot of people. Most of the income growth is concentrated on people who own companies, or own land, or who are very high skilled workers, and it’s not flowing to the great mass of people as it once did, so we have to fix that.
And we have to fix that because people won’t be satisfied if that sort of division continues. We also have to fix it because, as you mentioned there, pretty significant macro-economic problems (arise) if that doesn’t get fixed. We end up in a world where we’re constantly facing a demand slump, and that’s a very nasty place to be. So how do you do that? How do you address that need for redistribution?
Well, you can try to boost the bargaining power that workers have with firms. I think that’s a solution that has worked in the past in industrial history. I think the risk you face is that there are trade-offs there that some of that bargaining power is achieved by excluding others. If you don’t have unions that are really encompassing, then you have a lot of people that are left out.
An alternative would be to try and tax the rich and hand the money to the poor. In a lot of ways that would probably be more efficient. But then you have these nasty societal debates about, what do people have to do to earn that redistribution? Do they need to work? Do they need to contribute to society in some other way? Are we only going to redistribute to people who look like us, or are the same nationality or tribe, religion as us?
These are all very difficult questions, and I kind of throw up my hands a little bit in the book. I don’t have the answer. I think it’s an answer that we’re only going to arrive at over time as a result of a long debate between different groups in society.
The project that we run on these issues mentioned these points that you just addressed – the question of taxation – but also, if the value is accrued to the robots or the machines, the question becomes, who actually owns the robots and the machines?
Some of the arguments that we’ve had is that if you have different forms of ownership of this kind of capital, be it worker’s ownership, or be it some sort of public ownership via special purpose vehicles or something like this, that would allow you to basically re-socialise some of these gains.
I think that’s right. I like that idea, I like it from both an economic and a philosophical point of view. In a way, you can think about advances in technology and human knowledge as sort of like discovering a new natural resource. If you are a country that’s suddenly found a massive amount of oil, you might use that to create a sovereign wealth fund that you would then use to either support the social safety net or make other investments, because it’s held in common, it’s common wealth. And so I think there’s a parallel there with technology.
The difficulty I think, again, comes in deciding who is entitled to a share of that ownership. If you’re socialising the gains, is that limited to citizens of the country, and then are any immigrant workers second-class citizens? If you don’t limit it, then suddenly you probably have social pressure to shut out immigrants, and then that leaves people on the outside of the country all the poorer. I like the idea. I think the challenge is in finding the right balance between redistribution, and in who is allowed to participate in society.
It raises the question on what kind of governance level do you actually put these vehicles. Is it a national vehicle, is it a European one on an EU level? And what do you do with the people who are not going to benefit from this?
But the involvement of public institutions in this could also incentivise a bit of direction in this. I was struck, for instance, when we talked to Tony Atkinson two weeks ago that he mentioned, “Wouldn’t it be nice if you have chef-less kitchens rather than driverless cars?” Because that is, to a lot of elderly people in particular, a much bigger social problem than just driving around. So do you see that, without interfering in the technology development itself, but do you see that governments or public institutions in one way or the other could have a hand in directing this kind of technological process towards the most pressing social needs?
I think so. I think we are rediscovering the idea that government can direct technical change, and that it’s not necessarily a bad thing when it does. I think this raises a few questions, a few policy questions, and one is, what really are the most effective ways to do these sorts of things? Chef-less kitchens is an interesting idea, it would help a lot of people in addition to the elderly, but is that the best use of public funds?
Or would it not be better, say, to tax people and to use some of that gain to employ people who would like to work to cook for those who are unable to cook for themselves? Just as a hypothetical. I think that’s one of the tricky things you have to balance.
And then the other is figuring out how to provide the robot chefs, or whatever, to those who need them. And that, again, gets into the question of, who belongs, who deserves these sorts of things?
I suppose one thing to note is that we’ve gone so little down this pathway, that probably there’s a lot we could do without actually running into too many difficult questions. We’ve so underinvested in a lot of public goods, public research, that perhaps we should just do a lot more of that, and then we’ll cross the bridge of the difficult questions when we come to them.
Yes. I don’t think we as humans will ever run out of ideas about worthwhile things to do. This links to probably the most prominent policy debate in the context of this discussion, which is should we have a basic income or rather something else? I personally favour a job guarantee because it basically decouples the payment of activity from the content of activity. If you look at things like care for the elderly, aging societies, but also things like art and culture, activities in local communities; if you found a way to incentivise that, these are all based on social capital, and they won’t be automated any time soon.
What is your view on this discussion about, what are the basic incomes, or some other form of job currency, that might be one way of addressing this?
Well, I think you’re right that the basic income creates some fundamental problems along those lines. Jobs play several different roles, jobs created by the market have an important role in creating things and allocating purchasing power, but they also help structure people’s days and provide them with meaning and identity in their lives, and a sense of agency. They also provide workers with a kind of social currency that says that they’re allowed to participate in society because they’re contributing.
So if you use a basic income to solve the first problem, or to take care of the first role, you’re still left with the other two, and I just think that there’s no way that society will, at least not for some time to come, be okay with leaving those other roles unfulfilled. I suspect that the direction we will move in first will be less along the lines of a basic income programme or steps in that direction, and more along the lines of what can we do to make sure that everyone has employment, and how can we make sure that employment pays people enough?
That sounds to me like a job guarantee combined with wage subsidies, and I would expect that to be the direction for the next few decades, until technology perhaps means that that’s no longer something that works as a model.
When some people, they hear the term ‘job guarantee’, some have the wrong conception about this by assuming that this should just maintain the wage labour dependency, or maintain bullshit jobs. But basically what it just does is, it just gets rid of the market mechanism in the way this operates by decoupling the payment from the job with the content. It just basically then is down to local communities to decide what kind of social activity they would like to incentivise.
And of course the basic income, if it’s a universal basic income rather than a participatory income that people like Tony Atkinson support, you just spend a lot of money on people who effectively don’t need it, and just hope that you get some of it back via the tax system. But the tax system, if you look at recent performance, I think you will be forgiven for thinking that you won’t see much of that initial spending back.
Yes. There are a lot of issues there. I think the more generous you make it, the more taxation you have to do, especially if it’s universal, and that begins to be a very big lift, and you start chasing, especially in a world of mobile capital, you start chasing a lot of the money out of the system entirely.
It’s an interesting philosophical question: what should we want people to do? I think different societies will be more comfortable or less comfortable with what they feel should be asked of others in terms of what kinds of jobs they ought to do, what level of involvement – time spent on the job – ought to be, and how exactly those things should be incentivised, whether through something like social pressure, or through financial incentives, or through something like a mandate.
The only thing we can be sure of is that different countries are going to come up with different approaches to this, and I suppose we’ll see what works, and not everything will work in every context. I don’t know. In a way, if you’re a social scientist, it’s fascinating and it’s exciting, as times of great change usually are.
And of course there are lots of knock-on effects. If this kind of transformation hits the world of work, if you look at the welfare system for instance and how they’re financed. You have taxpayer systems like the one in the UK that might be less vulnerable; but if you look at insurance-based systems, like the one in Germany for instance, where there’s parity financing by employers and employees, and this is all dependent on what is traditionally conceived of as a normal job. But if normal jobs are becoming more scarce, you have a serious knock-on effect on how you finance your welfare system.
That’s right, that’s absolutely right. And you see this in the US as well, so much of what the welfare state provides is tied up with the idea that people, most people will have a fulltime contract job, and increasingly that’s not the case. Countries are struggling to rebuild the welfare state. I think one of the difficult things, and one of the things I try to point out in the book is that there’s a limit to how forward-looking people can be in reforming the social safety nets. There’s almost inevitably a backward-looking, or a lag involved, that the problems have to be so well-established as to enable the creation of powerful interest groups in order to get something done, in order to get change made.
I think, unfortunately, we’re at the stage of the process where the strains are beginning to come clear, but we’re only starting to see mobilisation of the political forces that might eventually yield change. So I think we’re a long way away from the kinds of social reforms we really would want to see to make societies equitable and well-off in this new digital world.
Yes, and Ryan, final question: if you were a policymaker, what would be your top two or three priorities? What would you try to look at and try to shape in the short-term?
Well, do I have a magic wand? Am I able to do whatever I like?
Just assume you can do whatever you want.
Okay. I think there’s a lot of low-hanging fruit, actually, in terms of dealing with the challenges we’re facing, and so I’d probably begin with increasing wage subsidies and taking demand much more seriously than rich countries have. There are an awful lot of ills that could be fixed by tight labour markets.
And then I would focus very heavily on public investment, and seeing how far that could take us. If we were to fund infrastructure, if we were to fund investment in industries that would create good jobs while also boosting economic growth, and funding basic research. Those aren’t radical policy ideas, but we have a pretty good sense that they would have some good effects. If you can kick the can down the road a bit, that’s maybe not such a bad thing.
Right, Ryan Avent, I’ve got a suspicion that this might not be the last book you will write on this subject.
I hope not.
Okay, thank you very much for talking to us today.
Thanks for having me, it was great.
This contribution is part of our project on the future of work and the digital revolution.
Ryan Avent is economics columnist at The Economist.