With every technological advancement since the dawn of time, conventional wisdom has warned that technology and automation kills jobs. But robots aren’t the root cause of our problems. Although technology has always changed the nature of work, this week’s guests Heidi Shierholz and Daron Acemoglu argue that there is no evidence that it has led or will lead to overall increased joblessness, unemployment, or wage stagnation.
Heidi Shierholz is a Senior Economist and the Director of Policy at the Economic Policy Institute. She was a Chief Economist at the U.S. Department of Labor under President Obama from 2014 to 2017. Her research and insights on labor and employment policy, the effects of automation on the labor market, wage stagnation, inequality, and many other topics routinely shape policy proposals and inform economic news coverage.
Daron Acemoglu is a Professor of Economics at the Massachusetts Institute of Technology. He is the co-author of the New York Times and Wall Street Journal bestselling book ‘Why Nations Fail’, with James A. Robinson. In 2005, he received the John Bates Clark Medal, awarded to economists under forty judged to have made the most significant contribution to economic thought and knowledge.
The zombie robot argument lurches on (EPI): https://www.epi.org/publication/the-zombie-robot-argument-lurches-on-there-is-no-evidence-that-automation-leads-to-joblessness-or-inequality/
How robots became a scapegoat for the destruction of the working class (The Week): https://theweek.com/articles/837759/how-robots-became-scapegoat-destruction-working-class
Automation, Job Loss, and the Welfare State (Council on Foreign Relations): https://www.cfr.org/event/automation-job-loss-and-welfare-state
Robots, or automation, are not the problem (EPI): https://www.epi.org/publication/robots-or-automation-are-not-the-problem-too-little-worker-power-is/
Robots kill jobs. But they create jobs, too. (Brookings): https://www.brookings.edu/blog/up-front/2019/03/18/robots-kill-jobs-but-they-create-jobs-too/
Where Do Good Jobs Come From? (Project Syndicate): https://www.project-syndicate.org/commentary/automation-vs-job-creation-by-daron-acemoglu-2019-04?barrier=accesspaylog
The Revolution Need Not Be Automated (Project Syndicate): https://www.project-syndicate.org/commentary/ai-automation-labor-productivity-by-daron-acemoglu-and-pascual-restrepo-2019-03?barrier=accesspaylog
Annie Fadely: Hi. I’m Annie Fadely. I’m a senior policy associate at Civic Ventures. We’re going to get to the episode in a second. But first, I wanted to tell you about an amazing new show that we’re listening to, called Your Primary Playlist. If you’re like us, you might feel a little fatigued already, by this presidential primary. With such a diverse field of candidates that’s running in 2020, it can be super difficult for even the most tuned in to keep track of who’s saying what, and who we agree with.
Well, good news, that’s where Your Primary Playlist comes in. It’s an interview style show. It’s hosted by Emily Tisch Sussman, and it breaks down the major issues facing our country, featuring all women guests. The real experts behind the issues like gun safety, healthcare, and climate. Upcoming episodes include a conversation about the future of the Democratic party, with speaker Nancy Pelosi, a discussion of reproductive rights, with Cecile Richards, and an interview with Neera Tanden, the president of the Center for American Progress, all about the economy. You can find Your Primary Playlist wherever you get your podcasts, so do yourself a favor, and subscribe.
Here’s the show.
Nick Hanauer: We cannot end up in a circumstance where we fear automation.
Heidi S: Many of us may lose our jobs because of robots.
Nick Hanauer: It sounds like a newfangled problem, but the solution is actually pretty old school, taking care of your workers.
Paul Constant: Yes. It’s not that complicated.
Annie Fadely: From the offices of Civic Ventures in downtown Seattle, this is Pitchfork Economics, with Nick Hanauer, Confessions of an American Capitalist Caught on Tape.
Nick Hanauer: I’m Nick Hanauer, founder of civic ventures.
Paul Constant: I’m Paul Constant and I’m a writer at civic ventures.
Nick Hanauer: So Paul, in this episode of Pitchfork Economics, we’re going to talk about automation and jobs. And that’s a subject that’s very much on the minds of lots and lots of people.
Paul Constant: Yeah. It’s something that is always discussed in the economic sphere and there’s always a little bit of a panic to it. There always has been a little bit of panic to it, the discussion of robots taking your jobs.
Nick Hanauer: And what’s interesting is there’s a lot of panic in the newspapers today, but there’s been panic in the newspapers for a really long time. Annie in our office actually did a favor for me and went back in history to try, and find headlines from newspapers. And it was astounding what she found going way, way, way back to the 1920s, 1930s. But a robot is after your job.
New Technology isn’t a panacea, New York Times 1980. It goes way, way back. And then 1955 New York Times automation is seen as boon danger. Economist tells join group, it could increase wealth and eliminate jobs. It’s the same headline, it’s just different styles in the decade.
Paul Constant: So automation for sure is a really important force in human economies, but it isn’t a new thing and it isn’t even a thing that dates back to the industrial revolution. It’s a thing that dates back to the first time that a person tied a rock to a stick. And that my favorite example is digging holes. In the beginning, people dug holes with their fingernails and then somebody invented the shovel.
If you think about the shovel as a piece of automation, I mean, there simply no doubt that a person with one good shovel can dig literally a hundred times as many holes as a person with their fingernails. And so you could argue that that one shovel eliminate to a hundred jobs or something like that, or 99 jobs. And yet we continue to dig holes. There’s lots of people with shovels and shovels were replaced, of course, by back holes with [inaudible 00:04:08]. Even more holes and bigger holes and better holes.
Nick Hanauer: The tools it start with simple tools and then the tools get more complex. And with the complexity comes more complex problems and, better solutions and more complex problem.
Paul Constant: That’s right. And the really interesting thing about the transformation the technology brings and automation brings is that to build… To have a shovel requires an insane amount of technology. Just to make a shovel, it turns out to be a very, very hard thing to do.
Nick Hanauer: I couldn’t do.
Paul Constant: Nor could any of our listeners because it requires you to find the metal or smelt it and turn it into a shovel and then somehow attach it to a stick. No single person could do that. It takes an enormous amount of technology and enormous amount of cooperative labor to produce the shovel that enables you to dig holes so quickly to say nothing of building a backhoes. Which we’ll put a zillion shoveled diggers out of work, but requires an almost, just an incredible number of people with incredibly specialized skills and an incredible amount of technology at their disposal to build one of those backhoes.
And so the thing about technology is that it does replace tasks and jobs, but by so doing creates new jobs and new tasks in the process. And the question is how do we organize a society that changes in this way, in ways that benefit the broad society over time.
Nick Hanauer: So I mean, I want to be clear here that, that we’re not saying that automation is a made up threat or anything. I think the thing that changes with each new sort of automation panic you see in the news is the kind of people who are affected by it, the jobs that are affected by it, right? Like in the last century, agriculture went from a third of all jobs to 1% of jobs in the last century. And now what we’re seeing is automation in sectors that previously haven’t seen these levels. There is information and business services, healthcare and retail. So those are job sectors that exploded over the last century. And now they’re seeing a new form of automation. So it’s a real problem.
Paul Constant: Yeah. It’s a very big problem in the near term to find ways of transitioning from one sort of technological regime to another technological regime and defined ways to integrate the folks who’s tasks or jobs are being automated away into new tasks and jobs. But over the long term [inaudible 00:07:16] one of my strong feelings is we’re not going to run out of jobs until we run out of problems. And we’ve got a lot of problems to solve. I don’t think we’re anywhere near to the end.
Nick Hanauer: So the automation is not the problem. It is the displacement. That’s the problem.
Paul Constant: Yeah correct.
Nick Hanauer: I’m super excited to get to talk to my old friend Heidi Schierholz, who is director of Policy at the Economic Policy Institute but was formerly Chief Economist at the US Department of Labor under President Obama. Heidi and I have known each other for a long time and we have collaborated on some cool things, including the overtime protections rural at the Labor Department under President Obama. And we did good work together, but they did it too late. And then the Trump administration killed it. Sad, sad for everybody. I’m fascinated by her research on the effect of automation on the labor market, wage stagnation and inequality. She knows a ton about that stuff. So anyway, Heidi’s is super interesting and brilliant woman and it’ll be fun to talk to her.
Are you there?
Heidi S: I’m here. Yes.
Nick Hanauer: Can you hear me?
Heidi S: Is that Nick?
Nick Hanauer: It is, how are you?
Heidi S: I’m good. How are you doing?
Nick Hanauer: Can you hear me?
Heidi S: I can hear you perfectly.
Nick Hanauer: Awesome. And you are either joined by or soon to be joined by my colleague Paul Constant.
Heidi S: Cool.
Paul Constant: Hello.
Nick Hanauer: Can you hear Paul, okay?
Heidi S: Yes, me. This is Heidi and I can hear both of you.
Paul Constant: Okay.
Nick Hanauer: Hey, so let’s talk automation and jobs.
Heidi S: All right.
Nick Hanauer: That’s one of our favorite subjects and we’re super happy to have you here because you’ve put a ton of energy in your career. Studying the effects of automation on the labor market and wage stack nation and inequality and so on and so forth. So let’s just start with your generalized take on automation, are robots going to take all the jobs?
Heidi S: The answer to that in my view is a pretty clear no. And all we can appeal to is what has happened in the past. But we have had massive disruptions, massive disruptions, it’s like the industrial revolution, computers, electricity. Like all of these things have happened every single time. Everyone said, “Oh, this automation is going to displace all the jobs,” and it doesn’t happen. And it’s because, and we can talk about why, how this happens, but that the economy can absorb it and just transfers the resources from the increased productivity towards other things.
One of the, I think that sort of best take home example here is in the early 1800s, over 80% of the employment in this country was in agriculture. So if you were a futurist in 1810, you could have very legitimately been screaming from the rooftops that, “Oh my God, automation is going to take all the jobs,” And you would have been some very important sense been right because now less than 2% of workers in this country are employed in agriculture.
But you would have been utterly wrong in thinking that we’d have no jobs. We completely automated agriculture. But the spending that was freed up because of the increased productivity and agriculture went elsewhere and created jobs elsewhere. So no one thinks that that decline in agriculture jobs actually led to mass unemployment.
That doesn’t mean that some workers weren’t displaced. They were, that leads that can and does and will continue to lead to hardship for workers who do get displaced. But I don’t think that there is, there has never been any sustained meaningful decline in the overall number of jobs as a result of technological change. And I don’t expect that to happen in the future either.
Nick Hanauer: Yeah, I think you make such a good point. And I think in order to properly frame this policy debate, you do have to look at the historical trends. And I want to underscore one of the points you’re making here, which is, it’s not just that 80% of the people in the United States worked in agriculture. It’s also that in 1810 and I did have to look this up. I want you to know I didn’t know it. When you said it, there were only 7 million people who lived in the United States.
And today we have about 330 million people. And unemployment rates are as low today as they were when we had seven. So with 50 times as many people we are employing essentially all of the people or the vast majority of people, even though the nation’s largest industry declined from 80% to about 1%. And so the scale of the change, which is possible in an industry which is essentially automated out of… It’s automated, it’s really you can’t get to a bigger scale than that. And yet today yhere we are with essentially full employment with only 1%of workers devoted to agriculture. It’s really fascinating.
Heidi S: Yeah. Yet it has never caused, technological change has never caused any sustained meaningful job loss. That’s just empirically true. And it’s not true just in the US. It’s also true around the world. So that’s empirically true in the past, the only thing that they could appeal to is that, oh well it’s going to be totally different.
Paul Constant: But this time is different.
Heidi S: This time is different and you have always repeated all the time, this time is different, this time is different. I think that the thing that’s echoes of is that it is a diversion from real problems that are hitting us right now, that are hitting us in the near term. The medium term and the longterm and we really need to be focused on, but if you say, at some point we’ll have no jobs, then that sort of captures people’s imagination and it gets this really breathless conversation going and divert such important energy away from really, really pressing issues.
Paul Constant: Right. But it is understandable that people worry about this because the thing about the human mind is that it definitely weighs risks more heavily than rewards. And it weighs what it can see more heavily than what it is forced to imagine. And you can definitely see with your own eyes how certain kinds of innovation will take either a task or a job away, but it is profoundly difficult to imagine what will replace it, right? Like just can’t know.
Heidi S: One of the things that I think is important when people think about how automation changes the labor market going forward is the pace of change actually really matters. And so there’s one industry, the autonomous vehicle industry that I don’t know, I only, I mean the site of couple of things, but they’re only from what I just googled. I’m not an expert in this area, but definitely an area where people tend to think that change is coming quickly and it’s going to be really dramatic.
But when you think about what that means, so according to Google, there are around 2 million commercial tractor trailers on the road right now. They each cost in the ballpark of $150,000, so that’s $300 billion of non autonomous tractor trailers out there right now. So it’s not like somebody’s going to come buy up all $300 billion worth of totally functional fine tractor and non autonomous tractor trailers and just trash them and replace them with much more expensive autonomous trucks overnight.
I mean, even in this industry where I have no reason to believe that change isn’t going to be coming soon. You can still see that on the ground that that change is going to happen slowly or relative, even in that industry, relatively slowly. And that’s one of the ways these things get absorbed without having a lot of actual economic harm some of, to many of the workers that were in those jobs, yet people retiring all the time, people transferring and changing jobs all the time. And then there’s just, you could sort of see how there’ll be fewer and fewer jobs in that area over time.
Nick Hanauer: The, the macro economic scale is very useful, it’s very informative. But you know just this week Amazon announced that they were enacting a new automated shipping procedure in their warehouses. So what would you say to a worker who was concerned about losing their work right now, in the face of automation?
Heidi S: That’s a really good point. And this is one of the things that I’ve, a trap I fall into, many economists do. And we just like constantly talk about we need, we can’t let our focus be diverted from this thing. That’s not going to change the net number of jobs in the economy. But that doesn’t mean some workers won’t get displaced. It is absolutely true that even though I don’t think we’ll see a big decline in the overall or any decline in the overall number of jobs as a result of technological change, some workers who do jobs that become automated will lose their jobs.
And we actually don’t have things set up very well in this country for that situation. So if like the… I think of us sort of three legged stool of we need for people who get displaced for any reason, not from necessarily from automation, but from any reason. So we need like when people start talking about this often the first thing they jumped to is education and training, that’s really important. We can’t stop there, we also need to strengthen safety net.
So people have transition step that help soften the blow and get them transitioned until they can find another job that works for them. And then what I would say is the most important for the third leg of that stool is we need to do all of the things that this podcast has been talking about to increase good jobs. Because one of the reasons that it’s so hard for people when they lose a job due to automation is that the jobs that are available out there aren’t that great because wages have been held down for 40 years. We don’t have good benefits, all of the things that have eroded job quality. So a whole agenda of boosting worker power to get good jobs is actually sort of counter-intuitively really core towards helping people who do get displaced by automation.
Paul Constant: The point you’re making about the good jobs and the relationship between having an economy with lots of good jobs and the disruption created by technological unemployment or automation is a really profound one. Because obviously if folks whose jobs are being automated have a ton of great alternatives, well, that makes the poll problem go away, that’s overstating it, but definitely softens the difficulty of the transformation. And I think that’s a really cool point.
Heidi S: Not often thought of in this context, but it’s just so fundamental to what does it mean when you lose your job for whatever reason. If you are hanging on to the one union job like that or whatever, if you had a decent job and you lose that job and the labor market that you get dropped into just does not have good jobs in it, so way worse situation for you than if the labor market you get dropped into has strong labor standards, strong unionization. So workers, like you said, you can find a decent job. You made all the difference.
Nick Hanauer: Yes really, it’s interesting because, I mean, I haven’t studied this, but it strikes me that a big part of the disruption has been that folks who had good jobs that were either automated or outsourced and formerly robust manufacturing sector or whatever it is, people, folks who owned good solid middle class wages with benefits and those jobs went away. Wasn’t probably in many ways that they couldn’t find a job. Just that going from $30 an hour with healthcare to $7 and 25 cents without healthcare, for those workers didn’t seem like a legitimate alternative.
And in fact wasn’t because there was no way they were going to be able to probably pay the rent to the house that they had or whatever it was that that new job because it paid so poorly wasn’t an effective replacement for the one that they had. And again, as we’ve discussed many times on this podcast, that wasn’t because that new job they were offered somehow was profoundly less productive or interesting or important than the old one. It was that the industry or company that offered that job hadn’t, the workers hadn’t had the power to negotiate a reasonable split of the value created in that industry. That was the difference.
Heidi S: Yes. I totally agree. You put your finger on it, that’s the really important difference. The problem there was not the technology, not the technological change. The problem there was the things that meant that workers didn’t have power on the job. So it’s the decline of unionization that led to that situation. And it’s easy to point to the technological change as the thing that led to the displacement. But really at its core, the problem there is not the technological change. It is the erosion of worker power that meant that workers have good jobs.
Paul Constant: Correct.
Heidi S: And find good jobs when they need to.
Nick Hanauer: And one of the things that we’ve said before in the podcast that I think is really worth, just resurfacing is that, we cannot end up in a circumstance where we fear automation and technology because automation and technology are largely how human societies have improved their circumstances since the beginning of human civilization. And I don’t think we have to fear capitalism either, which is simply a way of organizing ourselves into groups to solve one another’s problems.
What we have to fear is neo-liberalism and ideological framework that concentrates the winnings and socializes the costs of these technologies in these transformations that there’s no economic reason why new jobs can’t be good jobs and service jobs among other things can’t be good jobs and if they were, the folks who had been displaced in the older industries of manufacturing and other things would have transitioned much more easily to service jobs that supplied enough income and benefits to allow them to maintain their lifestyle.
Heidi S: I absolutely agree. And I would make one sort of friendly amendment, and the one thing I’m keeping my eyes open for in this world of tech changing, technology is we’ve always had that changing technology in the workplace is I think rather than some wholesale displacement of people from jobs. The thing is for the keeping my eyes on is more anecdotal evidence of how technology, when you have a situation of huge power imbalances in the labor market as we do now following 40 years of erosion of labor standards and institutions.
When you have that the power and balance then technology can sometimes be used as a cudgel against workers. So the thing that I think of like immediately is the warehouse workers being asked to wear smartwatches like the hyper monitoring of workers that, that if those workers were unionized, that plant would not… Those plant owners would not…. Warehouse owners, sorry, would not be able to unilaterally just say, okay, all you workers are now going to wear smartwatches where we can hyper monitor, like we can tell if you’re taking an extra bathroom break.
We can tell if you’re congregating with other workers, maybe you’re organizing. So we will, we can sort of surreptitiously retaliate against you, fire you or give you bad shifts or whatever it is. Like those kinds of technological changes and changes in how work gets done in that sense. But those kinds of things I think we do need to keep a real eye out for in this period we’re in, where there’s such a huge imbalance of power between workers and employers.
Paul Constant: So let’s talk for a minute about solutions. What are some of the best policies you know of right now that, that helped deal with that balance between automation and worker power?
Heidi S: What I think of the power relationship in the labor market, what we know is that employer’s sort of inherently have power in the relationship. They’re the ones who have the job with many workers. So each worker is vis-a-vis their employer at the power imbalance is strong there. And then we used to have these things, these standards and institutions that provided workers with countervailing power. So there was more balance in the labor market and those were the years where we had, as productivity rose, we also saw wages of low and middle income workers rise.
Over the last four to five decades, we’ve seen those things erode dramatically. So now we have this huge, now we can see what unchecked employer power looks like. And its stagnant wage that’s rising inequality. And so it’s the things that reverse that or what we need and the chief among them, our policies to boost unionization, policies to boost labor standards and institutions. Those are key, and we could go on like things to, I mean like even macro economic policy to keep the unemployment rate low is a super important piece of worker bargaining power and on and on. It’s this suite of things that will shift power back to workers.
There’s lots of… I feel like discussion around this conversation about automation where people who have benefited from the rigged economy often want to sort of make people think that, oh, it’s unfortunate that we’re seeing this rising inequality. But it’s just the natural outgrowth of a modern economy with lots of technology and that’s just flat out wrong. What the situation we have now are the results of choices that we have made. And we could do this in another way.
Nick Hanauer: One thing we’d like to ask all of our guests is why do you do the work that you do? What brings you to this field and to this job specifically?
Heidi S: Oh, that’s a really good question. I grew up in Iowa, I did not have like a political upbringing and then I sort of got around, read some interesting stuff in college and got around people who are reading and talking about more interesting stuff and just gradually became aware there was something sort of the kind of unveiling over the decades of my adulthood of just seeing just how kind of rigged the economy is against regular people and it’s almost… For me it’s been once you see that there’s no turning back. There’s no, like this is a real problem. It’s very fundamental. I have some skills to bring to this. And that’s the end of the story. This is something I wanted to be at, sort of a career well-spent to work on these issues. And so those are the… It’s kind of a moral drive for me.
Nick Hanauer: We are so grateful for you having spent the time with us on our podcast.
Heidi S: It was super fun.
Nick Hanauer: Thanks so much. And say hi to Larry in the gang for us. And we’ll talk soon.
Heidi S: Okay cool. Have a good one.
Nick Hanauer: Take care.
Heidi S: Bye.
Nick Hanauer: Bye.
Heidi S: You too bye.
Nick Hanauer: So one of the things that struck me from our talk with Heidi is the historical component because the terminology used to describe robots taking your jobs is always so modern. It’s always cutting edge technology that’s taking away job or there’s always some new segments showing off how cool this new self driving car, so something like that. So it feels like a problem that’s never been confronted before. I think looking at it in historical context is provides a whole bunch of new insights that people don’t ordinarily think of.
Paul Constant: That’s right. And we have to be super attentive I think to the real challenge that individual workers have and in particular industries around automation and transformation and make sure that we have a policy framework that supports those transformations and make sure that people are okay through those transformations. But we definitely don’t want to fear technology. We don’t want to fear automation and, we don’t want to be bullied by the trickle downers into believing that we should keep wages low to keep the automation at bay. That’s just that a con job to keep wages low and profit high.
Nick Hanauer: So it sounds like a new fangled problem, but the solution is actually pretty old school taking-
Paul Constant: -care of your workers.
Nick Hanauer: Yes. It’s not that complicated. So I’m incredibly excited to get to talk to our next guest Daron Acemoglu who is one of the foremost economists really in the world. He teaches at MIT, he is the author of one of our favorite books, Why Nations Fail. And has done a ton of research around jobs in automation. Well, listen, thank you again so much for joining us today. We are huge fans of your work and your book, Why Nations Fail, is required reading in our shop. It’s just such really kind of a breakthrough book on why some societies do well and why some don’t. So have you continued that line of research?
Daron Acemoglu: Sure. Actually, we have a new book coming out in September called, The Narrow Corridor, which is about the evolution of liberty over the ages and today. So it sort of continues, but sort of adds to the themes of why nations fail. So let’s see whether you’ll like that, or you’ll hate it. You’ll see.
Nick Hanauer: I love it, I love it. Oh Great, we look forward to reading that for sure. So mostly we wanted to talk to you today about your research on jobs and automation. And you are one of the world’s leading experts in this field. So maybe lay out your view of the issues and the problems and how we should think about it.
Daron Acemoglu: Well, I mean, first of all, just to lay the scene, I think automation together with the sort of changes in the institutional order we’re facing in the U.S and around the world and climate change are the three most important problems facing us today. And what’s common between all three of these problems? I know we’re going to focus on information, but just to plug for the rest, also is that I think they are urgent. They’re challenging, but they’re also opportunities. There are opportunities for us to actually up the game and do new things and also in the process strengthen our society. But those opportunities are very difficult for us to find the right sort of political and social consensus to support those solutions. So that’s exactly my sort of vision for what’s going on with automation.
Also, automation broadly, including industrial robots, AI, and all sorts of AI supporting applications such as big data, machine learning. They are providing amazing tools for us to increase productivity and diversify the set of activities we can perform. But at the same time, they are also creating great strains for our existing institutions that undergird shed prosperity because they destroy jobs. And often they are also required new ways of creating jobs that we are not completely familiar or completely comfortable with yet.
And they also change the ways in which the balance between capital and labor between states and societies are being configured. So all of these are potentially discombobulating for society and make it much harder for us to maintain the same sets of sort of solutions that we have devised over the last century for maintaining inclusive prosperity, prosperity that benefits society at large.
Nick Hanauer: In Why Nations Fail You, you talked about how, and I’ve thought this was really surprising how people in power in some nations oppose the industrial revolution because it seems to me in modern times that the people in power are the ones who are promoting automation. And AI and things like that. Is my perception of it now true or do you see, are people in power sort of, is there a friction between them and the incoming waves of automation?
Daron Acemoglu: This question is telling about a number of issues that we should really talk about. One of them is you’re absolutely right. If you look at the history of technology and economic growth, one of the weaknesses, almost the Achilles heel of authoritarian extractive regimes was that they either blocked new technologies or they did not provide the right ecosystem for those technologies to be invested in and to flourish. So as you pointed out many countries opposed factory systems, railways, they were late in adopting many of the new technologies that brought greater productivity.
But there isn’t a sort of an iron law that says extractive regimes will always oppose industrial production or new technologies. In fact, many of the countries in Europe when the industrial revolution started were not fully inclusive and they actually had a vision of building factories that would increase productivity but would certainly not create shared prosperity.
So factory technology spread in the UK, which was the leader quite rapidly at the end of the 18th century and the beginning of the 19th century. But it was also based on the exploitation of women’s and children’s labor. It was based on sort of employing low skill, low wage labor and sometimes even employers wanted to use their repression and were successful not just against sort of low skill workers but against middle skilled workers.
But all of these changed in the UK and later in France and Germany as for example, there was more and more pressure from worker organizations, from citizens organizations for both political and economic rights. So it’s not a surprise that for example, movements like the charter movement sort of married economic demands together with political demands because they realize the two went hand in hand. So that’s how sort of more inclusive political and economic systems emerged.
Today when you look at it, it’s much more similar to the second type of response that you see among authoritarian regimes. It’s not the case that China for example, just to pick one, for this example, is likely to block new technologies anytime soon, but it wants to use new technologies in a way that’s not necessarily consistent with shared prosperity and certainly not consistent with rights and a voice for the majority of its population.
And even in the United States, I think there is a real soul searching and political tension about how we should use these new technologies for developing a new social consensus. So the one vision is, we’re going to need less and less workers and the very creative, the very rich are going to get rich and the rest are not going to share in this sort of prosperity. We’ll either keep them happy because we give them universal basic income. We keep them happy because they have TV shops or really repressed them or whatever. But this isn’t going to be a very much a two class society. I don’t think that’s the best way of proceeding and a different vision is to try through the more difficult thing is to see how we can increase productivity, use technologies. But in a way that’s consistent with shared prosperity.
Paul Constant: Yeah. So Daron, you’ve written that technology can either be used to displace labor or enhanced labor productivity.
Daron Acemoglu: Yeah, exactly, that’s where we’re coming.
Paul Constant: So can you explain what you mean by that?
Daron Acemoglu: And that’s exactly what I was referring to in an oblique way into previous, which is that, why is it that we have these two ways of sort of founding a new society, a new social compact, so to speak. But it’s not because like we’re going to do the same things no matter what. And then the only question is how much we’re going to tax high incomes. It’s not just about redistribution, of course redistribution is part of it. But actually it’s a small part of it. The bigger part is how we use the technology, what types of jobs we create in what direction do we push on the technology frontier? And the crucial observation here is exactly the one that you emphasize.
A technology platform, and almost all of the really transformative technogg creies we’re talking about are platforms that can be developed in a multitude of ways. All technology platforms can be used more for automating meaning for using machines, for tasks previously performed by labor, or they can be used for creating new tasks for labor and housing, labors productivity in both the sets of tasks that it was producing before and in new tasks that it can perform. And it’s that tension that’s going to determine where do we go.
And just to emphasize, automation is not a new thing. This tension has been with us. It started with industrial revolution. What was the industrial revolution? Industrial Revolution was perhaps still despite what we are undergoing right now, still the biggest phase of automation in the history of humanity, which is that people found ways of machines, ways for machines to perform the tasks that were previously done by skilled weavers, artisans, metal workers, black summits, and so on.
So it was all of this sort of automation process that was at the root of the early phases of the British industrial revolution. But how more shared prosperity came about wasn’t simply that, the British state introduced redistributive taxation. It did so, but very slowly and very very late, much more it came because there were concerted moves for using technology for other things. A whole host of new occupations from factory floor work to managerial design work later for non production work both in the manufacturing and non manufacturing sectors were introduced.
And this was those sectors that both contributed to productivity and soaked up all of the labor that was increasing very much because agricultural labor was declining because of mechanization of agriculture. Population was growing, people were moving to major cities. But all of these people found jobs, or most of them found jobs ultimately because of these new tasks that were being created, because of these labor enhancing labor in stating technological functions.
Nick Hanauer: I just want to emphasize though, that the technology and automation has been with us for much longer than the industrial revolution.
Daron Acemoglu: Oh yeah.
Nick Hanauer: I mean, technology is in automation. I’ve been with us since the first person tied a stick to Iraq.
Daron Acemoglu: That’s right, absolutely.
Nick Hanauer: And then-
Daron Acemoglu: And fire is an amazing technological breakthrough.
Nick Hanauer: Right, absolutely.
Daron Acemoglu: Many of these political breakthroughs when you think about them, have both labor augmenting or labor enhancing and automation functions. Technologies broadening our horizon both proactively and socially. And that I think is what’s missing right now. So again, if you look at our recent economic history, since World War II. It’s a really interesting break in the middle of it. So for the first four decades after the end of World War II. You see fairly rapid automation in the U.S, but this is going handed head with other technological changes so that at the same rate as labor is being displaced by automation is being reinstated by other technological changes.
A lot of it is coming out of the, in the non manufacturing sector, but some of it is coming in the manufacturing sector. As a result, you will see the labor share for example, which means how much over the value added created an industry is going to labor as opposed to say capital. That’s more or less constant in many, many industries in the aggregate. But then around the 1980s, there’s a break. So you see the labor shared, especially in manufacturing, mining, and a few other sectors dropped very sharply. And you’re seeing a lot of automation again, mostly in manufacturing more recently now in services and other sectors, transport is coming.
And no countervailing reinstatement, no countervailing new tasks, new activities. So it’s like almost a very sharp break. And of course that break, if you look at it concise with two other big macro trends, we’re just stopped growing and productivity growth actually becomes lackluster. So that’s one the most puzzling things that the optimists of technology and people who are excited about Silicon Valley and the tech world revolutionizing everything don’t emphasize or find ways of avoiding is that we’re just not getting enough productivity growth out of these new technologies. And I think that’s related to how we are not using them the right way.
Nick Hanauer: That’s right. But Daron, there is a very simple explanation for this break, which is power, the declining power of working people to fairly share in the value created by all of these technologies. And that leads to a bunch of undesirable outcomes. Using a very, very simple example. If you use a technology to replace 10% of your workforce, there are all sorts of ways in which that result can express itself in the economy. One of those ways is where the owners get all of the benefits of that labor saving. Another way is where everybody who continues to work for that company gets a 10% wage increase. Right?
Daron Acemoglu: Yeah absolutely.
Nick Hanauer: In the latter case when workers are now being paid 10% more. One of the knock on effects of course, is that they’re buying 10% more things and when they buy 10% more things that’s having a feedback effect in the economy, which is requiring businesses to make 10% more things, which is creating jobs both in existing industries and in adjacent industries. And when we allowed the power dynamics in our economy change and productivity and wages decoupled in the early seventies.
Aggregate amounts of spending declined his income concentrated at the top and productivity fell as it was no longer a crucial for employers to continue to invest, to save on labor. And so all of these things are inextricably intertwined. And I guess that my basic point is that we… It is the social norms that define how technology is deployed and who benefits not anything intrinsic to the technology itself.
Daron Acemoglu: So I completely agree, they are intricately intertwined, but I think this interconnected nature of technology and institutions power social norms is really the thing to focus on. Because there’s one narrative which I think is too simple, which is in the 1980s, that was a shift against organized labor and capital became stronger and that’s explains everything. I think that is just not a good enough explanation. So the shift happened and under Reagan, and perhaps before there was anti-union movement, federal minimum wages were not increased or were increased very little from then on.
But I think that by itself explains a small part of what’s going on, and the really way that this whole change itself is intertwined with technology, if you want to think about it somewhat differently, is that because technology, as I emphasize, is a platform. How we develop it depends on social norms, depends on the situation, will depends on power or dynamics. And if you did develop technology and more of the automation will pay, that changes the power dynamics, and these to become self reinforcing.
Nick Hanauer: That’s pretty sure.
Daron Acemoglu: But crucially is really the changes in technology that happen. And it’s not just organized labor, it’s broader than that. I’ll come back to them in a second. But it’s changes in this technology shaped by our social preferences and social organizations that then have this major effects on distribution of resources, the labor share wages and productivity. So for example, I think we can try to sort of gauge the effects of workers get poorer and therefore they don’t demand certain types of products that’s there.
Some people will emphasize, but I think at the end that’s not a big deal. The bigger deal is really whether we are using technology the right way. I think that’s where it is. And that’s where all of these social institutional factors really are filtered through technology. And there I think, it’s not just the effects of organized labor, it’s really a broader sort of shift. And I would count among the factor that really mattered there, three, that are particularly important in addition to declining power of organized labor. First government’s role in shepherding technology, especially into more creative, riskier, more blue sky type of projects has really declined after the end of the cold part and a variety of other things. So the amount of government support and government leadership in technology has declined.
Second, I think it is also absolutely critical that we have developed an ecosystem of technology around the tech world where the mindset is geared almost completely towards automation. So if you think of the most successful companies and which are the sort of the source of all of the… or most of the research and AI machine learning automation type activities, it’s all about getting the foldable human out of the picture. Have a business model that doesn’t depend on humans, get a business model that’s very cost effective, get a business model where you can have high degree of automation because that gives you greater control, greater flexibility, greater lift, nimbleness and all of this sort of ecosystem also means that technology is flowing more and more towards automation at the risk of doing other things with technology that’s going to increase the productivity of labor.
And third, and equally important, there has also been a change in social norms and institutions about how we approach firms and the corporate world. So what is sometimes called the shareholder’s values revolution meant that, it became accepted for companies to ignore all of their stakeholders, including their workers, their customers and suppliers and just focus on shareholders, and in particular large shareholders, which meant and much more ruthless way of cost cutting breaking promises to workers, getting rid of workers at the expense of benefiting from tax breaks and, taking jobs to where you can find new loopholes to make it less costly in terms of taxes or other things.
So all of this sort of ensemble of things that have happened because of social reasons have added to this bias picture in which technology has been excessively directed towards automation and not for creating good jobs, high wage jobs for middle-class workers.
Nick Hanauer: Yeah. So I think I’m massively with you, but I do want to push back on just one thing you said, which is that-
Daron Acemoglu: Sure, of course.
Nick Hanauer: The power dynamic isn’t as significant as people think because if you simply held the median family in America harmless for the last 40 years of rising inequality, instead of earning $59,000 a year, they would earn about 105 or $110,000 a year.
Daron Acemoglu: Absolutely, yeah.
Nick Hanauer: And if you were to think about the kind of economy we would have in which-
Daron Acemoglu: It would be very different. But I want to clarify what I meant. What I meant is that a world in which, midian family is empowered technologically, socially and economically compared to this one is not going to be very different. Of course, it will be the very different. What I meant it, keep the organized power of union. The power of organized labor which means the trade unions power, the same level as in 1970s. But everything else the same. Government leadership and technology disappears.
We have a shareholder’s values, revolution [crosstalk 00:53:13], and Silicon Valley defines what automation is going to do. We wouldn’t be anywhere better. And why not? Because actually we know from Europe, and from cross country, some of this in my work where you have more unionization and more wage push. The adoption of automation technologies, if anything is faster. So if you called all the other social equilibrium costs and then you just sort of make the truckers more bargain faster more and, machinists bargain more, what are you going to do? Is you’re going to just make the adoption of the automation technologies even more powerful and more pronounced and nothing much will change. You really need the whole social equilibrium for change. And that’s why I was emphasizing the other things. And then I think the decline in the power of labor is very important, but within this context,
Nick Hanauer: But that being said, in 1965, it was unambiguously true that the median family in America was more prosperous than any other median family on planet earth. And if you roll forward 40 years, that is unambiguously not true that the median family in Sweden or Denmark or Switzerland or Germany is more prosperous than the median family in the United States. And that’s not because they have less automation or less technology, it’s because their workers have more power.
Daron Acemoglu: But it’s again, it’s not like just workers’ power in a sort of an organized labor way.
Nick Hanauer: No, no, no, no.
Daron Acemoglu: And one way of understanding that is actually compare.
Nick Hanauer: [inaudible 00:54:58].
Daron Acemoglu: Is compare Germany and Sweden, which you pointed out to say Italy or Greece.
Nick Hanauer: Yes.
Daron Acemoglu: So in Italy, in Greece, labor is actually very powerful. There is quite militant unions that can hold their own, but it’s not actually helping workers all that much.
Nick Hanauer: Correct.
Daron Acemoglu: The reason why that’s the case is that actually, and France is actually closer to Italy and Greece also. And the reason is because these unions have not really created an environment in which they can work with firms in a way that’s going to encourage the adoption and development of technologies that are going to increase labors prospects. And as a result, what happens is that you have actually faster elimination of workers using technology because employers don’t want to deal with these workers.
So the genius of the German and the Swedish systems is, and I will always be in a little bit sort of ambiguous about this in the past, but I’ve become convinced that any shortcoming of the system is much less than its benefits. Is there corporatist structure, which means that unions are protecting the workers, but they are doing so in a way that’s actually consistent with longterm or middle term planning of businesses for their survival and prosperity.
And it’s that combination that has, for example, made German companies adopt automation technologies, for example, robotics very rapidly to compensate for their aging workforce, but at the same create and a lot of new manufacturing jobs. That’s also the system that has enabled the Swedish economy to sort of adapt to this changing environment.
And again, that’s sort of a very different view than say, the shareholder’s values, we’re going to just look after the interests of the shareholders or what union is saying. We’re just going to ask for the maximum wage gain that we want and we’re going to have a fairly conflictual relationship with the employers. And that’s why it’s sort of that, has to be embedded in a social equilibrium where there is the right kind of communication between labor and businesses.
Paul Constant: Yeah. So interesting. So we have gone over our time, but we have one last question to ask you, which is-
Daron Acemoglu: Sure, of course.
Paul Constant: Why do you do your work? What brought you to it?
Daron Acemoglu: Oh, I mean, I’m excited about it. It’s the perfect, I couldn’t have dreamt for a better job, which is, I work on things that I think matter for the world and it’s fun. It’s new, it’s sort of has a degree of creativity, it’s a degree of communication. I started by saying, I think the three most urgent problems facing us, our automation technologies, how we use them, institutions, how we shape the institutions for the future climate change. Those are the topics I have the freedom to work on. Those are the topics I do my research on. I have a passion for it and it’s fun to do it. So I think it’s the happiest medium for me.
Paul Constant: It’s so great. Well, Daron, thank you so much for spending time with us talking about these thoughts.
Daron Acemoglu: Oh my pleasure, thank you for having me.
Paul Constant: And I can’t wait to read the new book.
Nick Hanauer: Yeah, we can’t wait to read the new book and I know that. And I know that-
Paul Constant: And I’ll come back, if you want to have another one after the new book comes out.
Nick Hanauer: We will get absolutely 100%. Okay. Well thank you so much.
Daron Acemoglu: Thank you Nick, thank you Paul. Great talking to you.
Nick Hanauer: Okay bye bye.
Daron Acemoglu: Bye bye.
Nick Hanauer: So Paul, those were fascinating conversations about jobs and automation and we were really lucky to get to talk to those folks about it. What do you think?
Daron Acemoglu: I feel a little a little better about it just in terms of having identified the problems and the groups that are affected and, looking at it from sort of an evolutionary perspective as opposed to a terror. It’s good to know that there are such smart people who are thinking about this problem. And I think that’s the kind of perspective we need as we think about dealing with a displacement in years to come.
Nick Hanauer: And as I reflect, I continue to believe that we need not fear automation or technology. Because that’s how human societies have always improved our welfare. And we need not fear capitalism, more markets because that’s the institutional arrangement that allows people to come together and use technology to solve human problems. What we have to fear is an ideological framework like neo-liberalism, which creates a circumstance within which technology and markets concentrate the winnings and few hands and socializes the losses on everyone else. That’s the problem, is it technology can be used to benefit everyone. Or it can be used to benefit the few and harm that many.
And that has nothing to do with the technology itself really, it has to do with the social norms, democratic regulation and institutional frameworks that we use to manage the technology. And so, it all does at the end of the day. As your friend, JW Mason says, “It’s all techs all the way down.” It always comes down to who gets what and why. And we need to have citizens who demand that the technology be used to benefit the society broadly and not be lead in into this idea that, well, like, what are we going to do? It’s technology. There’s not, that just because those people at Facebook are essentially monetizing humanity’s deepest vulnerabilities well that’s fine.
It’s a free country. They should be able to do that. No, no, like we have a right to decide how technology is used in our society and on our society. And I think that’s for me the main takeaway.
Paul Constant: Yeah. And I think that that both Heidi and Daron had sort of the same conclusion that you need to have policies and, a philosophy that puts more people at the table and includes everyone because if you take a strictly extractive point of view and all the burgers are flipped by robots, then the robots aren’t going to buy the burgers.
Nick Hanauer: Who is going to buy the burgers? So it’s for sure.
Paul Constant: It’s just super important to make sure that everyone is benefiting in some way from these amazing technological marvels that we are creating.
Nick Hanauer: On the next episode of Pitchfork Economics, we are going to talk about this thing we called educationism with a Diane Ravitch, one of the foremost experts on the subject.
Annie Fadely: Pitchfork Economics is produced by Civic Ventures. The magic happens in Seattle in partnership with Larj media, that’s L-A-R-J media and the Young Turks network. Find us on Twitter and Facebook at Civic Action. Follow our writing on medium at @civicskunkworks and peek behind the podcast scenes on Instagram @pitchforkeconomics. And one more, you should definitely follow Nick on Twitter at @Nick Hanauer. As always, a big thank you to our guests and thanks to you for listening from our team at Civic Ventures, Nick Hanauer, Zach Silk, Jasmin Weaver, Jessyn Farrell, Stephanie Ervin, David Goldstein, Paul Constant, Stephen Paolini and Annie Fadely. See you next week.