The Robots May Be Coming, But Probably Not for Your Job

Most discussions about AI and the future of work tend to go in one of two directions: either excitement for a ‘post-work’ utopia, or alarm over the end of work. On this episode Mark talks with Aaron Benanav, an economic historian, postdoctoral researcher at Humboldt University of Berlin, and author of ‘Automation and the Future of Work,’ about why this whole debate sort of misses the point.

You can learn more about and purchase Aaron's book here.


[MUSIC PLAYING] MARK BLYTH: Hello, I'm Mark Blyth, I'm the Director of the Rhodes Center for International Economics and Finance at Brown University. This is the Rhodes Center Podcast. On this episode, I talk with Aaron Benanav, he's an economic historian and also a Post-Doctoral Researcher at Humboldt University in Berlin. Over the past few years he's become an important voice and conversations around AI, automation, and work in the 21st century.

Most discussions around AI in the future of work tend to go in one of two directions, either excitement for a post-work utopia, or alarm over well, the end of work. Aaron's new book, Automation And The Future of Work upends this debate in a fascinating and useful way by placing automation in a global context.

Guess what? Manufacturing jobs are disappearing, but this is almost nothing to do with robots. It's a fascinating book, and we had a great chat, and I hope you enjoy it. Aaron welcome to the podcast.

AARON BENANAV: Thanks for having me, I'm glad to be here.

MARK BLYTH: So let me jump right in, and for people who haven't read the book yet, the probably most people who was not in the podcast, let's give them an overview of this. Now, you can talk about what the tech people say. Well, the automation fetishists talk about.

But as a straightforward economist, I'll put my straightforward economist up. The way I understand this and the way most people understand this is there's a thing called skill biased technical change. So the basic story is that manufacturing output is actually up, because productivity is up, because we're much more productive because capital substitutes for labor at the margin really efficiently in manufacturing not so much in services. So that explains why we've got more and more stuff. But at the same time we've got fewer and fewer workers.

Add to this the fact that there's a kind of a skills premium on wages, and I can explain inequality at the same time. So what's wrong with that story because lots of people seem to like it?

AARON BENANAV: That's a great question. And I think there's a number of different ways to approach it. We can talk about what's going on, especially in low wage markets, and where inequality is happening in the labor market. I think that sort of out of character with the skill biased technical change account, a lot more inequality is happening within the occupational categories rather than across them. And I think that's not something that's really captured by the skill biased technical change account.

But I think that the larger story there has to do with what's driving the loss of jobs in manufacturing. Is it productivity growth? And it's very common to point out that it's certainly true that manufacturing in a country like the United States, it's still true that we're producing more and more manufactured goods even though highly labor intensive jobs are obviously shipping overseas. It's not the case that the US produces less in volume terms than it did in the past.

So, for that reason, a lot of people are saying, look it's really a productivity story that's driving this, and precisely of the kind that you were saying. But when you look at the numbers and you look at productivity growth over time, although it's true that manufacturing today we produce more than we did in the past, the rate of growth of manufacturing, the rate of growth of output, or the rate of growth of the volume of industrial production is really slowed down significantly.

And so I think that that story, that big driver declining growth, and manufacturing declining growth in the economy as a whole is having much broader effects on the labor market than the skill biased technical change story captures. And I think we've seen that now the broader effects on labor market. We're really seeing that even as the so-called college wage premium has stabilized, as college age workers with skills are seeing stagnant wages as well as less educated workers.

MARK BLYTH: So let's play with the idea that essentially outputs, but it's not up as much as you think. And then the two other variables in the equation you point to in the book are productivity. Right? And then what's the last one?

AARON BENANAV: Employment.

MARK BLYTH: Exactly. So what is the relationship between the three of them then? What's the story you're telling is different with those three variables?

AARON BENANAV: Yeah, so it's really a very simple equation, and it's actually the one that Adam Smith begins wealth of nations with. He says, look, there's two causes of the rising wealth of nations. One, is the increase in the number of workers that are employed productively in the economy. And the other one is how efficiently those workers work. And that's just the simple idea. We can break down the growth rate of the economy into these two components productivity growth and employment growth.

And so especially in manufacturing I think, what we've seen is that there's a technology story that really suggests that the main drivers of change are technological in nature. It's a kind of productivity led story of employment loss. So, as productivity growth rises, if output growth rates are stable, more and more manufacturing jobs are going to go away.

But that story just doesn't make sense, not only does it not make sense in the US, it doesn't make sense of the global story. And deindustrialization is happening not only in the United States, and not only in Europe, it's also happening across quickly developing countries like South Korea, it's happening South Africa, Brazil. Mexico has been deindustrializing now for decades. And in all of those places what we see is that this deindustrialization process is not driven by rising rates of productivity growth, which is what the technology story would have us believe, but rather falling rates of output growth.

And it's precisely because this is a global story. It even affects China and India, and according to the UN and entire world since Twenty-Thirteen. Those are the kind of things that are pointing me towards a story in which what's really going on here is industrial overcapacity. It's really a world market story, it's a story about globalization, and it's a story about how as more and more countries are trying to get into industry and are forced to try to make places for themselves in global manufacturing.

We've seen more and more trade redundancies rather than trade complementarities, very hard for countries and companies to find places in the world market and that's driving down rates of growth across the world. Those countries like China that did grow very quickly for a while actually grew quickly because they were able to take market share away from other countries and it's kind of make their deindustrialization worse.

MARK BLYTH: So what we have at a macro level then is the classic Marxian problem of overaccumulation leads to underconsumption, leads to falling wages, and then you get caught into a spiral. And sort of the standard critique of that, because in a sense I'm giving it by describing the Marxian problem as underconsumption, was a kind of Keynesian remedy, which was, well, if you got too much supply, just ramp up the demand. So we haven't been able to do that. But I'm just thinking about that structurally, not just because of COVID spending and stuff, but the bottom 2/5 of the world really want to grow.

This is in a sense one of the issues I have with the notion of doughnut economics. You know the thing where you meant to like bring people down from the top and push them off to the bottom. Well, those people who are at the bottom like they really want to get quite far up, and that means the consumption of more and more stuff.

Now put in the environmental damage that that would cause to one side. That would suggest that there is demand in the world economy, particularly in so-called emerging markets, the shift to Asia, bare or demography.

So how much of this is a kind of a temporary thing to do with aging Atlantic economies? And is this a problem that has its own solution in the sense that structural forces in the world economy as it shifts as we move towards Asia, towards younger populations? You're going to see that consumption picking up, or is that just naive?

AARON BENANAV: No, I think that's a really great question. I guess there's multiple parts there, right? So I guess what I add to a classic Marxist account sort of blending the story you told about over accumulation and overproduction and underconsumption to a story about a transition from manufacturing and industrial-based economies to service-based ones. So I add a Baumol cost disease element to that story.

The transition of services is not only happening in the west, we're actually just in the past year or two, we passed the point globally where 50% of the world's workforce is said to be in services. And the slow growth of productivity that we see in the western services is actually much worse and a lot of lower and middle income countries, where many of the people and services are in the informal sector, and they're doing really low productivity work. There are many countries in which productivity growth in services is actually negative, not merely low.

And so those are obviously drags on global productivity growth rates. And they kind of as manufacturing is becoming more impacted as it were, by overproduction, and we're seeing more of the global population move in to services, I think that's going to be a drag on growth. I do think that the big problem with the story about how-- it's really true that obviously, poor people in the Global-South need more stuff and more income. I found Thomas Piketty's revision to the elephant graph really convincing that the growth in the past 40 years or so, has gone much more to the top 1% of income earners rather than helping those at the bottom.

MARK BLYTH: So maintenant Piketty there, I mean, there's a question of how many variables we need to explain this stuff. So, let me take a couple out there. Again, put on my streams economics heart for a minute and you mentioned Baumol and cost disease. So a more standard explanation is look, things are worse in the west because you've got bigger service sectors. You can't add capital to a haircut very easily. So you just get crop productivity leaving aside whether wages are really driven by productivity or not. It does kind of explain low productivity.

So, to what extent do we need to invoke a global cause, when a local cause which is essentially you get a big service sector your productivity is going to slow down? Again What's wrong with that story?

AARON BENANAV: So you know something that I want to think about that's related to that, which has been presented to me as a critique of the Brenner Account that I partially rely on, is that they say, look, there is a demand shift story here. Like there is a tendency in societies as they get wealthier to more of their income growth is going into services. And that's a major cause of this transition. And my response to that has always been to say, well look, even economists like Dani Rodrik are now emphasizing this premature deindustrialization problem, it's clear that this transition is happening in many societies that are very far from reaching that kind of income level that would cause that transition to take place.

it's really a story since the:

And so I think that that story about how industry as this major driver of growth has become less and less available to countries. And only countries with incredibly strong state capacity. Infrastructure educated populations like China have been able to do it in recent years, and their capacity to do that has really-- when globally manufacturing growth rates are very low for one country to achieve the kind of rapid industrial expansion that they need to industrialize, they have to be taking market share away from other countries. So we've moved more towards a zero sum growth situation and industry as compared to the past when I think there was more shared growth in the 50s and 60s.

So it's like on the one hand, you need this story about why manufacturing is becoming congested and is no longer an engine of growth, and then marry that to a story in which well, yes, it's true in already wealthy countries this has been a problem and it might have that demand shift character at some level, but that's adding to these other issues pushing the whole world onto this slower growth path.

I think it's very interesting to ask a question about what the role for public stimulus and public investment could be given that kind of structural story. I don't think that just telling a structural story gives us all the answers or invalidates policy options. So I'll just say that.

MARK BLYTH: Well, I mean, we can get right into that because that's at the core, and it gets us into agency and notions of what can be done about this. I mean, the depressing structural reading of your book would go something like this, when I'm very much carnal and Mark Herman Schwartz on this one. This is a Schwartz reading you, I'm going to try and do that one. Now so forgive me Mark if you hear this podcast.

But essentially they are going to be something like, there was this magical confluence of things that came about in the 50s. So basically, semiskilled labor and abundance dollar or supremacy for this technology. And the select few that had access to this had really above normal growth for a while, and then everybody tried it. And once you did it, you kind of get a fallacy of composition problem at the global level, and then the whole system starts to destabilize.

And some countries try one solution. China integrates with massive labor cost advantage. America basically financialization and digitalizes is another way of dealing with it. And unfortunately the only thing that really benefits working people was that for the system, and now that's gone you're kind of stuff. There's a long run mean reversal to a kind of croppy or capitalism with lower wages.

Then there's another version of this, which you spoke about in a talk and you alluded to, which is the kind of Kaldorian understanding that says, no, actually, demand is independent of supply, and if done properly, it can be productivity enhancing. You get these called evildoing effects. And that's a way in which you effectively build your way out.

So, thinking forward, I'm thinking about the carbon transition. You're going to need according to the latest estimates by McKinsey, which should always be taken with a grain of salt. I think the figure was something like $34 trillion. So Adam, too, said that was something in the order of was it 20% of GDP for several years for the European Union? That's a lot of cash. But that much cash can make a hell of a difference to the aggregate demand situation of most countries. So in a sense, can we solve this problem or is it structural? Should I go with the depressive reading of Benanav? Or should I go with a kind of hopeful Kaldorian reading of Benanav?

AARON BENANAV: I think it's a great question and you know I'm interested for history to decide. I think the best we can do is make our educated guesses about the situation and its possibilities. I think that there is likely to be certainly some major bump that's available through green technologies and through public investment. I think that the issue with those stories is that it's very hard to imagine an actual decoupling taking place, that is to say that we've really will significantly reduce our-- that we can simultaneously grow quickly, reduce our use of carbon, and transition to renewable energy sources.

It just seems a little bit tough to imagine that happening. And that's why I think that there are more radical versions of the Green New Deal that see a kind of majar investment in green technologies alongside reduction of the workweek, alongside transitions out of earlier and longer transitions into retirement, more educational opportunities for young people. Just basically ways to shrink the labor supply alongside increasing the demand for labor. And I think that insofar as you're going to imagine, yeah, a solution from above I think that that is probably what it would have to look like.

I just don't know that I really believe that we're going to see states coordinate in that way. I'm kind of pessimistic. I mean, look at the CORONA crisis, right? It's been incredible to see how poorly it's been handled especially at the international level, and now around vaccine distribution. So I think that like it seems much more likely to me that just as we saw in Twenty-Nineteen and Twenty-Eighteen and in the US in Twenty-Twenty, we're going to also be seeing a lot of social movements and a lot of people on the streets making radical demands around what that project of transition looks like.

And to my mind, I think that one of the big limits of Keynesian demand management in the 50s and 60s, was how technocratic it was. And I think that precisely the kind of movements we've seen in the streets which really don't trust elites to do the right thing, that those kind of movements are going to play a really big role in the story, and that in a lot of ways that's where I put my faith and hope and energy is in the pressure for transformation that those movements could provide.

MARK BLYTH: So I'm with you in the technocracy taken to its extreme gives you populism because you hollow democracy and you hollow effective capacity of citizenship. So I'm with you on that one. But the other side of it, and your book is a little bit lighter on this than some takes it a route there, but it does hearken to a solution that basically says, hey, we should work less and which it is. And in a sense aren't you asking people to be poor? It's not a really tough sell.

nds me of a big debate in the:

I think we've been down that path before, and we've seen that it really doesn't work. And that direct provision of services to the poor to actually directly improve their livelihoods is going to be a much more effective solution to raising their standard of living than hoping that by inflating the entire economy and getting really rapid growth, some of that will end up benefiting poor people. I think we've seen enough of that to suggest that we need to try a different path or a different solution.

ree times as big as it was in:

Tell me what's wrong with the race against the machine thing. I have my own criticisms, but I'm particularly interested in getting yours out because I think they're particularly telling.

AARON BENANAV: I mean, I guess that in a lot of ways. There's two approaches to answering that question. One is just empirical, like I just think that if you look at the numbers and you look at investment, you just don't see the kind of story that they're telling. Like the story that they're telling would imply that if we were living in an age of brilliant technologies and incredible transformations, you would expect the economy to just be growing very quickly but not absorbing very much labor. And in that case, the problem we would face would be a pure distributional problem, like we could just tax profits and give it to workers and be done with it, and enjoy a post work life.

And it just doesn't correspond to the data, it doesn't explain this secular stagnation pattern, it doesn't explain the fact that productivity growth rates have actually been, especially from the perspective of the economy as a whole really, really, really abysmal. In fact, the decade that they say is the decade of the most amazing transformation of the twenty-tens was the worst decade in the post-war period for investment and it was the worst decade for productivity growth. So I just think that it's too much of a technological story, a technology driven story that doesn't have an independent social and economic dimension to it.

But what I found-- the reason why I ended up focusing on automation, I should say as well, there's a story about the limits of it as a technology story itself that I think has been coming out just very recently about the limits of the current capacities we have for artificial intelligence, just aren't panning out in the way that people thought that they would, which implies that essentially these systems are going to be good about 95% of the time. But that means that you're not really going to be able to depend on them for truly critical work. There's going be a lot of human involvement and work for a very, very long time.

But the other side of my interest in the automation account, is just that it has this utopianism to it. Has this kind of attempt like, as you were saying the account that I give is pretty pessimistic and structural, it's a really like a kind of long-term. Societies have grown old and decadent kind of account in some way. And I really like about the automation story that it's very optimistic. And it's trying to come up with really innovative and original policy solutions. Because I think that in a way, if you think automation is going to get rid of work, you think wow, society has to change in a dramatic way in order to adapt to this.

It's very similar to the way that when people take climate change seriously they say, wow, societies is really going to have to change now as a result of this. And those are the kind of things that I find interesting and compelling. And I think their effort to build a world where work is in the center of people's identities, and where people are secure just for being human beings like UBI, Universal Basic Income has this hope for a kind of post scarcity society that I find very compelling.

But because they hitch their wagons so much to technology, I think because that story isn't going to pan out. We have to provide other bases for optimism. Other kind of Social Stories about what we can do to make society better that are not so technologically deterministic.

MARK BLYTH: I'm going to read something that came across the wires when we were doing our talk today, and it's from a guy called Mark Lockwood, who lives in the United Kingdom, and he's a tech worker of [INAUDIBLE] and who will give up too much. He says, just from a public perception point of view, advanced technology and never really resulted in good news for average earners. Tech either replaces you, get you to do more work and often for less pay. And any of your work have you come across an instance where a recent implementation of advanced tech led to a significant improvement in pay and conditions at a base level. I accept the theory that you outline, but the public mistrust that's out there isn't likely to shift until there's a perceived real local benefit.

So I just want to focus on that for a minute and say, isn't that also a problem with techno optimism in the sense that it assumes that hey, wouldn't it be great if like everyone in California is sad home and played crap guitar. If we think of a COVID, I mean, one of the problems that people have been quite rightly pointing out is social isolation. That jobs are so much more than just the work. They are at least in principle, social networks and tides that bind and all that sort of stuff.

And along with that techno-optimism, is a kind of hyper individualism that goes along with that. I'm pretty pessimistic about. And yet what you're telling me is that honestly, we're running out of those jobs. And this goes way beyond your book, but the final chapter speaks to about, how do we find social meaning? How do we find communities in a world where workers under threat from all these different forces?

AARON BENANAV: Yeah, that's a very interesting question. And I think that in some ways. Although I find these proposals equally limited, I think that they're very interesting in answering a question. Some people like Daniel Susskind have said, UBI isn't enough, because UBI is just meeting people's demands, basically it's meeting their animal needs. It's kind of a way that exactly like, if you give people a little bit of money, they can pay their rent, they can sit at home, they can play video games or watch Netflix, and it's a really terrible life.

And part of that is definitely a story about technology and how it's being implemented. I think people are becoming aware of the way that social media via Facebook and Instagram and also things like Netflix isn't the only way these technologies could be developed. And I can give you a more positive story about the possibilities of technology. But directly related to what you said, the proposal Susskind makes is for something like a basic capital grant.

And to say that people need to not only have a means of meeting their needs, they have to have the means to shape the society that they live, and to be part of the story of-- having access to capital means having that opportunity to reshape the basic infrastructure in some sense. Maybe I'm being a little hyperbolic, but to shape the productive capacities of society within which we move.

And so he says, look, people have to have control over that. And I think that that starts to point toward a broader idea that it's only when people have I mean this applies as much in the individual workplace as it does to society. It's when people have more of a say, there's more democratization of the investment function so that people can actually begin to shape the social world that they live in and make it more amenable to social life.

I mean for me, moving from COVID America to COVID Berlin, was an incredible transformation just because there was a lot more public life here, and there's a lot more people outside. There's so many public spaces. And I think people are maybe, because of the welfare state a little bit protected from workaholism and like just working all the time, and it just meant that there was more public life, and I think that that's something that America clearly really needs, but also something that we need much more broadly.

MARK BLYTH: Well, I could go on, you could go on, but podcasts are best when they don't go on too much. I hope that we've whetted everyone's appetite to go out and get your book and read it. It's a great book, great contribution, and really thought-provoking.

Thank you very much for joining us on the podcast.

AARON BENANAV: Thank you so much for having me. It's been great talking to you.

MARK BLYTH: This episode of the Rhodes Center Podcast was produced by Dan Richards. For more information, go to watson.brand.edu.rhodes. Thanks for listening.


About the Podcast

Show artwork for The Rhodes Center Podcast with Mark Blyth
The Rhodes Center Podcast with Mark Blyth
A podcast from the Rhodes Center, hosted by political economist Mark Blyth.

About your host

Profile picture for Mark Blyth

Mark Blyth

Host, Rhodes Center Podcast