(The following is a transcript of episode 54 of the Review the Future podcast.)
Ted Kupper: I think most of our listeners are probably familiar with the concept of technological unemployment, and they’ve probably heard us going on about it considerably. But let’s just very quickly, in case anybody is joining us fresh, talk to them about the basic thesis of your new book Rise of the Robots.
Martin Ford: The basic thesis is the same as my earlier book, which is that technology, machines, computers, robots, artificial intelligence are essentially going to displace people more and more. We’re getting very close to the endgame, where technology is really able to do a lot of things that previously would have been confined to human beings. We’re really entering a new age there.
Jon Perry: This topic often is scoffed at by economists, which is something you mentioned in your first book The Lights in the Tunnel, that it was almost unthinkable to a lot of people. I’m curious, as that book came out in 2009 and now six years have gone by, how has the conversation around this issue changed since then?
Martin: Well, I think it’s definitely changed and it’s become a lot more visible. As you say, when I wrote that book, I mentioned that it was almost unthinkable in terms of the way economists approached it, and that’s less true today. There are definitely some economists out there, at least a few, that are now talking seriously about this. So, it’s really been quite a dramatic change as these technologies and the implications of them have become more visible.
Having said that, I still think this is very much an idea that is outside the mainstream, and a lot of the economists that do talk about it tend to take what I would call a very conservative tack, which is they still tend to believe that the solution to all this is more education and training–all we have to do is train people so that they can climb the skills ladder and keep ahead of the machines. I think that’s an idea that pretty much has run out of steam, and we probably need to look at more radical proposals going forward. But definitely it is a much more visible topic now than it was back in 2009.
Ted: We had a guest on this show recently, Mark Lewis, who’s a professor at Trinity University at San Antonio, and he very forcefully made the argument that education is just not going to get us out of here, and he had some pretty compelling numbers.
One thing I wanted to ask you about was there’s been a review of your book written that I read by the economist from George Mason, Robin Hanson. Of course, he is somebody who is somewhat skeptical of this. But I think despite the somewhat dismissive tone of the review, there were a few things that were interesting enough to bring up and give you a chance to respond to.
One thing that I thought was particularly interesting is that he brings up the history that despite a long time now of technological improvement accelerating at an extremely fast rate, we haven’t seen a similar rate in what looks like job displacement, or job loss, as a result of that. He makes an assertion in the review that we don’t have any particular reason to think that contrary to prior experience, a big clump of displaceable jobs lies nearby ahead.
I was wondering what your response was to that. Do you think we do have any reason to think that jobs are going to get displaced soon that haven’t been able to be displaced in the recent past?
Martin: Historically, he makes a valid point. Clearly, computers have been getting better, faster exponentially literally for decades now, and clearly we haven’t seen an exponential impact on the labor market. If we had seen that, then there wouldn’t be any jobs left already. So, his observation is correct.
But it’s also important to note that the nature of an exponential is that, at first, of course it has very little impact. But eventually when you get further into it, it begins to have a very dramatic impact. I think that we’re getting to the point where we’re certainly going to see a much bigger impact than we have in the past, and it’s really about machines and computers finally approaching the point where they’re able to really substitute for a lot of things that people used to do on a much more broad basis than has been the case in the past.
The people that tend to be skeptical of this, they will always point to history, and there certainly are examples, going back up to 200 years. I mean, you can go all the way back to the luddite revolt in England, where workers went on a rampage, destroying mechanical looms and things like that. But it is true that this has come up again and again since then, and it has always proved to be a false alarm so far. I think that has to do with the fact that, in the past, the technologies were relatively primitive, they were often mechanical in nature, and, of course, they weren’t very broad-based, they were often specialized.
The classic example of that is agriculture. This is the thing that the skeptics will usually point to first. It used to be in the late 1800s that actually nearly everyone in the United States worked on a farm. Now, the number of people in agriculture is less than 2%. So, all those jobs basically went away–that was millions and millions of jobs. And yet clearly it didn’t have a negative impact at all, it had a positive impact. People went to different areas, they found more meaningful jobs that paid better, and we also have much cheaper food now, at least relative to our income, so overall we’re much better off as the result of the mechanization of agriculture.
So, skeptics will point to that and say the same thing is going to happen today, and I would push back against that and say, “Wait a minute, you need to think about this a little bit more,” because clearly agricultural technology was highly specialized. It was mechanical, it was tractors and machines that revolutionized farming, but those same technologies couldn’t be taken out of agriculture and be repurposed into the service sector. They weren’t going to flip hamburgers, and go into factories, and do all of the other jobs out there.
What we’ve seen historically is that these technologies have hit on a sector by sector basis. First, it was agriculture. Later, it was manufacturing; factories automated, and to some extent, offshored as well. People move from sector to sector–from agriculture to manufacturing, and now everyone is in the service sector. So, the really destructive thing I think we’re going to see in the future is as information technology really impacts the service sector, which is where everyone works.
But the key point is that information technology is a broad-based general purpose technology. It’s everywhere. It’s not specialized like agricultural technology, so it’s everywhere. What that means is that there isn’t any sector out there waiting to absorb all these workers, everything is getting hit simultaneously now. So for that reason, I really think it’s quite a different scenario than what we’ve seen in the past.
When you combine that general purpose nature of the technology together with the fact that you’ve got this exponential rate of advance, and also with the fact that for the first time machines and computers are really taking on cognitive tasks as well as muscle-power tasks–when you combine all of those things, what you end up with I think is something that is a utility, kind of like electricity, but instead of just delivering electric power, it’s now delivering, to some extent, machine intelligence. It’s delivering the capability to solve problems, make decisions, and more importantly, to learn. You’ve actually now have got algorithms that can learn based on real-time and historical data, and machines can figure out how to do all kinds of jobs. That’s quite new and it’s quite disruptive.
Jon: In terms of new sectors though, do you think it’s too farfetched to think that within the service industry, which is a very large industry, that because of a human touch being a big portion of a lot of service jobs, that this might be an area where jobs are more resilient and that we might move toward more jobs that leverage human empathy or people who desire specifically a human in their service interaction?
Martin: I think, to some extent, that’s true, and I think also that people probably vastly overestimate that as being a mechanism that’s going to protect a lot of jobs. You can already see that computers and machines are getting better. You see articles in the press about emotional artificial intelligence and things like that, where machines are beginning to really push into that area.
You can also point to, again, historical examples. You can look at bank tellers, they really have a “people” job. If you go into a bank, they will be very friendly, they will smile at you, they will try to cross-sell you other things that are being sold by the bank. But that hasn’t stopped people from moving en masse to using ATMs. People, when they’re presented with automated convenience, they very often will make that choice. So, I think that that’s true to a certain extent.
Another issue is very often those jobs that involve a kind of human touch aren’t especially good jobs. You’ll hear a lot about home healthcare aid for the elderly and service occupations, like giving pedicures and things like that. But we can’t all have jobs where we just provide those kinds of personal services to each other. There has to be a core of employment where people have to have the kind of incomes that are necessary to support all of that. What is becoming increasingly clear is that technology is really having an impact on a pretty broad swath of those jobs.
Ted: I want to drill down on this idea of what particular clump of displaceable jobs are sitting in the future, and I do feel like there’s an answer to this in your writing, which is I’m seeing two big clumps of the service sector that seem in terrible danger of immediate displacement, one being that clump of the service sector that drives vehicles, and another being that clump of the service sector that basically processes natural language and does little else for a job.
So, basically people who are telephone customer service, for example, they’ve been fighting a wave of automation for a while. But I think the current technology is still considerably less good than a human–but not because it can’t search a database for an answer, it’s because they can’t understand you well enough. It does seem like there’s massive progress being demonstrated, at this point, on both of those fronts: the automated vehicle thing, which of course everyone is familiar with at this point, and also deep learning seems to be a major breakthrough for teaching computers language.
Martin: Yeah, I think you’re right, those are both two disruptive areas, and I would point out that automated vehicles is not just about displacing drivers–it’s much bigger than that. Google’s plan for automated vehicles in the future, their vision is that the cars would become a shared resource, at least in urban areas, so that you actually won’t own your car anymore. Instead, you’ll pull out your mobile device and you’ll call for a shared vehicle. That’s incredibly disruptive, if you sit down and think about what that would mean.
Think of all the car dealers, the independent repair shops, insurance agents.
Ted: Parking lots.
Martin: Parking lots. In Los Angeles, there are 10,000 people who work in car washes.
Ted: Yeah, we live in Los Angeles, so we’re well aware.
Martin: A lot of those people have criminal records and things like that. It would be difficult for them to find jobs elsewhere. And the reason they have jobs, of course, is that everyone owns their car and takes a lot of pride in it, and seems to believe that going to a hand car wash is better than going to an automatic car wash.
If cars someday become fleet-owned and you just call up and then the car arrives, then all of that basically goes away. So, it’s not just specifically jobs for people that drive vehicles, and in fact there are millions of those jobs, but it’s even thousands and thousands of independently-owned businesses, and insurance agencies, and everything like that. So, that’s hugely disruptive if that happens according to the kind of vision that Google, in particular, has for it.
Beyond those areas that you mentioned, another area I focus on is fast food. I think that it’s inevitable that we’re going to see more automation in fast food eventually. There’s a company in San Francisco called Momentum Machines that’s working specifically on building a machine that can crank out about 400 hamburgers an hour. In many ways, fast food, at least the production side of it, more like just in time manufacturing than a service sector job. If you’ve got factories in China automating to build precision electronics, then it stands to reason that we can also eventually automate hamburgers, and lattes, and tacos, and stuff like that as well. I think eventually those jobs are also in the sights.
The other area is general retail, and there are a couple of things that are going to impact that. One is that online retailers like Amazon are going to continue to get bigger and continue to disrupt brick and mortar. And as that happens, you might think in theory that doesn’t mean that jobs will disappear, that jobs will just disappear from stores and then reappear in distribution centers owned by Amazon, but the reality is that those distribution centers are becoming increasingly automated. Amazon has made big investments in Kiva robots and in other forms of automation, so I think it’s reasonable to assume that as jobs move more and more to the online side, that they’re going to be more susceptible to automation.
Then the other thing is you’ve got people bringing their mobile devices into retail environments, and eventually you’ll be able to tap into something like IBM Watson that will give you real time assistance–real time natural language assistance–right inside the store, and will give you probably faster and more accurate information than you would be able to get by trying to hunt down an employee in the store.
Of course, there’s also actual robots increasingly in retail. There’s a hardware store in Silicon Valley that is actually introducing robots that can not only give you information but can actually lead you to the place in the store where you can find a particular item, and things like that.
So, all of that is going to have I think a pretty big disruption throughout the service sector. And then on top of that, you’ve got all kinds of white collar office jobs, really any kind of job where you’re sitting at a desk, in front of a computer, manipulating information in some relatively routine way–that’s going to be highly susceptible. You see examples already in terms of accounting, in terms of journalism, where you’ve got algorithms that can actually right articles pretty proficiently; in legal areas where you’ve got document processing. That’s just going to get bigger and bigger going forward.
Jon: Now, we’ve had information technology for not that long, but we’ve had it for a little while now, and there are certainly some troubling economic trends that we see–stagnating wages, rising inequality, for example. To what extent can we say that, say, information technology or even technology in general is at least partially the cause of these economic trends? And if that’s the thesis here, what would be the best way to actually try to tease that apart and measure technology’s impact on the labor force going forward?
Martin: It’s kind of a challenging problem. I believe obviously very strongly that information technology has been an important part of it. I would not argue that it’s all of it by any means. And in my new book, Rise of the Robots, I point out about seven general trends that you can look at.
That includes the fact that wages have stagnated while productivity has continued to increase, so productivity and incomes have kind of decoupled. It includes the fact that the share of income going to labor as opposed to capital has gone into a pretty precipitous decline, especially since the year 2000. The labor force participation rate, meaning the number of people who are actually actively engaged in work, is falling. We’re seeing wages for college graduates actually going into decline, so it’s not the case anymore that people with higher educations are doing extremely well; a lot of people with college degrees are also being impacted.
So, there are a number of things you could look at there, and the thing is that if you take any one of these and you look at the research that economists have done, there are lots of explanations. Technology is nearly always one of the explanations, but there, of course, are other explanations. There’s globalization, there’s a basic change in our politics, which will become more conservative. In particular, there’s the decimation of unions in the private sector. Depending on who’s doing the analysis and sometimes what their agenda is, they will point to those things as being more important than technology.
But what I believe is that if you take all of that evidence collectively, if you look at all of those things together, it’s really hard to come up with one explanation other than technology that can explain all of those things. And so for that reason, I think that technology over the last few decades has been really important.
But still, the main point that I make is that I’m not too interested in arguing about the past. I think that argument is going to continue basically forever. I mean, economists still argue about the great depression in the 1930s and what caused that; there still isn’t total agreement about that. I think the same will be true about the impact of technology over the last few decades.
My focus is really on the future, and one of the points I make is that a lot of these other trends, including to some extent globalization, and demographic changes, and, for example, that a lot of women have entered the workforce, and the fact that unions have been destroyed in the private sector–all of that is kind of played out, to some extent. It’s done with. I mean, we’ve already sort of reached a limit there.
Whereas technology, you continue to have this exponential increase. So, when you look at that, I think what it suggests is that going forward from this point, technology is really going to rise above all these other factors. It’s going to become increasingly clear that technology is really the thing that is shaping the future. I think that that will become pretty unambiguous over the next decade or two.
Ted: I think there’s a dearth of good statistical economic approaches to actually teasing this out, and certainly some people that I’ve seen writing, and they I’m sure have their own agendas, have argued relatively persuasively that, for example, most of the inequality increase that we’ve seen recently can be attributed to housing prices.
But I agree with your focus on the future. What do you think we should be looking for as markers going forward? Let’s assume for a moment that this thesis is correct, and technology has been some part of the equation in the past and it’s going to be a larger part of the equation moving forward. What should we expect to see? What would prove it right here?
Martin: I think that you could look to a number of things. I believe that we’ll see continuing increases in inequality, I think that we’ll continue to see stagnant wages, even as productivity increases. We may see increasing unemployment. So far, you have to say that, at least in terms of the way things are measured, the impact so far has not been so much in increasing unemployment per se, but in stagnant wages and in increasing what you might call underemployment, and less stable employment. That’s been the story so far. But it’s certainly possible at some point we could see real increasing unemployment.
Ted: And also it’s been reflected in a lowering labor force participation rate, right? Which unemployment, when the BLS calculates it, is as a percentage of the labor force. So, as the labor force decreases, unemployment goes down, even if no more people get jobs.
Martin: That’s right. A lot of people, if you give up looking for a job, and basically the way they do this is with a phone survey. So, if they call you up and you say, “No, I’m not looking because I’m discouraged, I’ve basically given up,” then you’re out of the labor force and you’re not counted as being unemployed. And yeah, we see a pretty dramatic decrease in the labor force participation rate. Again, some of that is demographic and the economists will squabble over what’s really going on there, but I certainly believe that some of it is technology, and it’s basically the reality that more and more of our population is essentially just not marketable in terms of the skills that they have and the capabilities that they have.
The other thing you’ve seen in conjunction with that is skyrocketing applications for social security disability, which is supposed to be just for if you suffer a serious injury on the job and can’t work anymore. But I think it’s pretty clear that it’s being used as kind of a last resort, almost like a guaranteed income. It’s actually the closest thing we have in this country to a guaranteed income program, is the social security disability program, if you can get in it. And people are using it that way because they don’t have any other options.
Jon: Let’s talk about guaranteed income, actually. My understanding in the new book is that you support a basic income. But in The Lights in the Tunnel, it seemed as if you were specifically not supporting that type of program. There’s a quote in which you say, “A program in which everyone is provided a relatively equal income in return for doing nothing provides no motivation for self-improvement, no sense of self-worth, and no hope for a better future. This is the problem with existing welfare programs.” So, if you support a basic income now, it seems that your perspective on that has changed?
Martin: No, I haven’t changed. Basically what I proposed in The Lights in the Tunnel was what I would call a guaranteed income, but it had incentives built into it. So, it was a guaranteed income where everyone would get something, but not equal, meaning that if you, say, graduate from high school, you ought to get a higher income than if you don’t graduate from high school. So, essentially what I really proposed was what you might call a modified version of a guaranteed income. But I still believe in giving everyone something.
But I think that it does raise a real issue. Let’s suppose that you’re a marginal high school student. You’re someone that really has to work at it in order to graduate from high school. And you know that when you turn 18, no matter what, you’re going to get a guaranteed income that’s the same as everyone else. I mean, that’s a pretty strong, perverse incentive for you to basically drop out of high school at that point, isn’t it? And I think that would be disastrous for society.
So, what I really have tried to do, and I basically offered the same essential solution in both books, in The Lights in the Tunnel and also in this new book, is I think that the idea of a basic income is good, but we need to build at least some rudimentary incentives into that so that we don’t really create an incentive for people just to not even educate themselves. That would be just disastrous if that happened. Because one thing I think that you can say is that in the future things are going to get more complex, and as a democracy, it’s going to be harder and harder for us to face up to these new challenges and figure out what to do. I mean, we need an educated population. We can’t afford to have a population that’s primarily high school dropouts or something.
Ted: Well, I certainly agree that we need an educated population. But I want to push back a little bit on this idea that a basic income would be disastrous for society, because I don’t think I agree with that.
First off, there’s intrinsic motivation in achieving higher status in society, and obviously I don’t think it’s going to change any time soon that graduating high school, graduating college is going to confer higher status on somebody. Plus, the reward of education is its own reward; people do it just for their own betterment, I think, if they’re given the opportunity. If you can go to school without being afraid you’re going to starve to death while you do so, it seems like that would encourage people to take an educational route over working a dead-end job and learning nothing, which is what a lot of people do nowadays if they don’t come from means.
So, I’m curious whether you’re basing that on just a personal belief or if you have some data that you think supports that. Because as far as I know, the limited studies that have been done have been pretty positive in terms of relatively low basic incomes not being particularly problematic with regard to people’s self-improvement.
Martin: Well first of all, I definitely wouldn’t say that I’m against a basic income. I think I’m for it. I’m just proposing a variation on it, and it’s not that dramatic a difference. You seem to believe that everyone should get exactly the same income, and I’m saying that there should be some levels there. They don’t necessarily have to be dramatic, and I’m not saying that the people at the bottom that get the lowest level of the guaranteed income should be impoverished or should get dramatically less than someone who graduates from high school. I’m just saying it would be good to have some basic incentives built into there.
Ted: Right. I guess I’m just attracted to the simplicity of a system that has no tiers, and therefore no bureaucratic overhead really, other than the one task of making sure everybody only gets it once and that they don’t collect two of them by fraud. Other than that, it seems like the administration cost could be extremely low, and if these tiers that you’re describing are not very dramatic, I’m also wondering whether they’ll be very useful as incentives.
Think about that marginal student that you’re wanting to encourage to finish high school. If he’s only going to get a little bit more than he would dropping out, it seems like you’d need basically an objective study to determine whether that was going to make any difference.
Martin: Yeah, it’s true. I would support studies into this. This is really just something I’m offering on a conceptual basis. Obviously no one has ever studied this. I mean, there have been pretty limited studies done on basic incomes themselves, and certainly the first step would be to have pilot programs to actually study this. I mean, the studies that have been done so far have been pretty limited.
Basically what I look at, in terms of proposing this, is the fact that we continue to have a really major problem with high school dropouts, especially in certain underperforming areas and so forth. I do think it’s really important to give people an incentive and also maybe to give them some sort of a ladder that they can potentially climb that’s obvious to them.
Ted: One thing I’d like to add to that, because I completely agree with you that we have a big problem of people not finishing high school in poorer areas of this country–it seems to me like one of the major motivations people give when they do surveys on this is that they are leaving school to go to work. I think if those people were entitled to a basic income of any kind, we might actually see that that would do a lot to get those people through school.
Martin: Yeah. The other points I make in the book supporting, again, a basic income–I don’t want you to get the impression that I don’t support that, because I do, and actually I devote a whole chapter to basically arguing for that–but one other thing is that I think there is what’s known as the “Peltzman Effect.” Sam Peltzman was an economist at the University of Chicago who studied the impact of safety devices put in automobiles, things like seat belts and so forth. What he found is that as cars got safer with more regulation, it didn’t actually improve the fatality rate. People still got killed in automobile accidents at the same rate.
The “Peltzman Effect” is also known as risk compensation, and basically what it says is that if you give people some kind of a safety net by making things safer, they will then take more risk in order to make up for that. You see the same thing with children’s playgrounds. If you ever go to a kid’s playground nowadays, you’ll notice that there’s all this spongy rubber on the floor and everything, and everything is this kind of low-key plastic, there’s nothing really dangerous anymore. And yet what they found is that the kids still break their arms at the same rate as they used to when there were high monkey bars and things. Apparently what’s happening is that the kids are taking more risks now that the playgrounds are safer.
So, the argument I make in the book is that if we had something like a basic income or a guaranteed income, you’d see that same kind of effect in the economic arena. People would have a safety net and they might be more inclined, for example, to start a business or leave a safe job in order to take a job at a small startup company that involves more risk and that type of thing. So it might actually, to some extent, create a more dynamic economy, so that could be a very positive effect.
If you give people an income floor below which they can’t fall, then many people would, I think, do things–entrepreneurial things–to actually increase their income above that, and that would create a pretty strong dynamic effect in the economy, which could be a very good thing provided that the incentives were correct so that you didn’t, for example, tax away all the extra income that people were able to make by doing something with a small business.
So, I believe strongly that that’s a very positive impact. But again, I go back to that incentive to get educated. A lot of these things, in terms of starting a business or doing other things in terms of the online economy, you need a basic threshold of education in order to do those things.
Jon: So, let’s zoom out a bit and talk about why we want the basic income in place in the first place. The obvious one is that people need to eat, of course. But in The Lights in the Tunnel, you talked a lot about the risk of consumer collapse, that if people don’t have money in their pockets to spend, that that leads possibly to a very bad cycle where businesses have nobody to sell to, the mass market falls apart to the point where businesses would fail, and eventually this would affect even the very wealthy as well, and that part of the reason that we need a policy of putting income into people’s pockets is just to keep the whole thing churning. Is that still a big emphasis in your new book?
Martin: Yes, I have a chapter on that. The new book is a little bit different from The Lights in the Tunnel. The Lights in the Tunnel was really kind of a thought experiment. I just basically sat down and kind of imagined a future and talked about some issues. There wasn’t a whole lot of data or specific examples in that book.
The new book, Rise of the Robots, has got a lot more actual economic data and specific examples.In the chapter where I talk about this issue, about the fact that basically we need consumers to buy the products and services that are produced by the economy, I actually do show some data that this is, I think, becoming a real concern, that you can actually point to actual evidence to suggest that this is really happening.
One thing that’s happening is that consumption is, like income, becoming much more unequal. It used to be, back in the 1990s, that the top 5% of households was responsible for about 27% of consumption. Now it’s almost 40%. So, fewer and fewer households are doing more and more of the spending in the US economy and it looks to me that that’s going to continue to be the case, and yet, at the same time, we see that, as automation begins to impact those higher level white collar jobs, even those households at the top of the income spectrum–and I’m talking about the top 5%, not the top .1%–but that top 5% is increasingly susceptible to this as well, so. It really becomes unclear how sustainable it is. You ultimately have to have enough people to buy what’s produced by the economy. It’s not just about the aggregate amount of dollars out there, it’s really about how those dollars are distributed.
You can think in terms of Warren Buffett or Bill Gates. They, in theory, have an infinite amount of purchasing power. They can buy anything they want. But, of course, they don’t want to buy anything–I mean, they’re not going to go out and buy a thousand cars or a thousand smart phones. They’re certainly not going to sit down and eat a thousand restaurant meals in one evening. So, they cannot keep the whole consumer economy going by themselves. So, when you take purchasing power from a thousand people and concentrate that into the hands of one person, which is kind of what’s happening, that has real implications. I mean, if you’re selling a product or service, it’s not just the total amount of dollars out there, it’s also unit demand. It’s “how are those purchasing power dollars distributed among consumers” that really matters, especially in the mass market economy that we have today. I think that, going forward, that’s going to be an increasingly a concern.
And you see some evidence of that already. In the retail business, what you see is that retailers that focus on luxury goods have been doing really well, and retailers that focus on the middle class have been really kind of struggling. You know, Walmart has had one earnings report after another where they have not been doing especially well, and it’s because their consumers basically don’t have any discretionary income, a lot of them are on food stamps.
So, the question is can you really have an economy that’s driven by people buying ferraris, and yachts, and expensive handbags? It’s just not sustainable over the long course… If you really look at all the big industries that really drive our economy, whether it’s automobiles, or consumer electronics, or financial services, they all rely on markets that are made up of millions and millions, and tens of millions of people, and we need to continue to have an economy where there’s reasonable distribution of purchasing power so that you literally have enough people out there to buy these products and services.
Ted: So, what is a reasonable distribution of purchasing power? Given the various measures that we have available to us, how high of an unemployment rate or how low of a labor force participation rate, or how low of a labor share of income would we have to get to, do you think, before we see serious repercussions in the sense that the market collapses or new types of consumer goods can’t make their market the way that they do now?
Martin: You know, I’m not an economist, so I don’t delve too much into making specific numerical projections. But I think even if you were to ask the economists a question like that, you’d get widely varying estimates. So, that’s a question that we’re just going to have to wait and see.
I think if you look at what’s happening already, there’s cause for concern. If you look back at the economic crisis that we just had, everyone knows that that was started initially by subprime mortgages, where people borrowed money that they couldn’t repay because they didn’t have the income. And the focus has always been on the fact that people were given mortgages they shouldn’t have been given, but it’s also important to note that the underlying cause is that these people simply didn’t have the income to support a middle class lifestyle really, and that’s a big part of what was going on.
So, when you have people that don’t have sufficient income to really aspire to the American dream or to a middle class lifestyle, that’s a real problem and they will tend to try to take on debt if they can in order to do that. And so, I think that there’s definitely a risk going forward that we could run into a similar problem. I mean, if incomes are not going to increase for most people while at the same time house costs continue to increase, healthcare continues to increase, educational costs continue to explode… I mean, if there’s no income, then the only alternative is really to take on more debt, and if people can do that, they will, and then we’re likely to get ourselves into another crisis.
Jon: You just mentioned the three areas that have been very resistant to lowering their price: healthcare, housing, and education. These are things that remain extremely expensive and, in many cases, have gotten more expensive while people’s wages have not gone up to compensate for that.
Since we’re projecting future technologies here that are going to eliminate jobs, should we also be projecting future technologies that might actually start to lower the cost of some of those things, like healthcare, and education, and maybe even housing? And is there a chance, that if we were to take a more optimistic view, that maybe a lower cost of living in the future might help us to avoid some of the more dangerous aspects of this future?
Martin: In the long term, I think maybe you can be hopeful about that. But in the near term, it’s difficult to see that. Again, in the new book, Rise of the Robots, I’ve got chapters on healthcare and on education, and I talk a bit about this. So far, it’s a real challenge. Information technology simply has not transformed those sectors the way it has other sectors in the economy.
In education, higher education in particular, you can look at what’s happening with massive online courses, MOOCs, and there is sort of an obvious path there that could lead to ultimate disruption and might really make a difference. But clearly it has not happened yet. But at least there, there is a fairly obvious path forward.
In healthcare, it’s even more challenging. I mean, you look at things like nursing, the need to take care of the elderly, and you can talk about building robots to do a lot of that, but it remains, as of now, a really daunting challenge. I mean, the idea that you can build an affordable robot that can do what a nurse does in terms of dexterity, and mobility, and problem-solving and so forth is really pretty far out there in science fiction land at the moment. We do have robots designed to help in elder care and areas like that, but they tend to be pretty specialized and they really are pretty limited in what they can really do. So, that’s an area that, so far, technology is really kind of falling short in the face of what is just a really important and dramatic challenge that’s looming demographically.
So, we can hope that in the longer term there will be big things happening there. I mean, we see also inroads, in terms of artificial intelligence, in medicine with things like IBM’s Watson that’s moving in to medicine and so forth, and there’s a hopeful story out there a little bit further into the future. But for right now, it’s not happening quite yet. So, I think that that will remain a challenge for people in terms of–people that work in other areas that are being disrupted, where automation is getting more and more advanced. And so their incomes are falling or their incomes are stagnant, and yet at the same time they’re seeing exploding costs in education and healthcare–that’s a big problem.
The other thing is a lot of these things are double-edged swords. I mean, we can hope for disruption in higher education, and that may make education more affordable. But at the same time, there are millions and millions of people that work in higher education; that’s really an area that provides good jobs for millions of people who have advanced degrees. If those jobs suddenly go away because of this disruption… There are two sides to that, clearly. Same thing with healthcare.
You mentioned housing as well. One thing that might, in the longer term, be disruptive in housing that might make housing more affordable is construction scale with 3D printers. I mean, there are experiments with those already and they could really make a big difference in terms of building affordable housing. But it would also eliminate huge numbers of construction jobs. So, these things have two sides to them, and it’s not clear exactly how that’s all going to play out.
Jon: So, eventually people are going to notice that this is happening, and what happens with that reaction. One thing that you mentioned, again, in your previous book, is that could lead to people being concerned about their future unemployment and therefore spending less, which could speed up the type of problems that we’re worried about. But it could also trigger the type of societal change that might get us out of this mess. When do you see people taking note of this, and what do you think happens as this starts to become more obvious to people?
Martin: Again, that’s really impossible to predict, but I do think it’s becoming more visible and I think, as you say, people are beginning to worry about it a bit more. I think that, in general, unemployment today is perceived differently than it was 20 or 30 years ago. I mean, it used to be that to be unemployed meant that you got laid off, and a lot of times the same company would hire you back. It was really perceived as a temporary things.
Nowadays, people understand that a lot of times if they lose their job, it may take a long time to find another job, and if they do find another job, it means a permanent reduction in earnings. So, I think it’s fair to say that the public in general and consumers in particular perceive unemployment and the threat of unemployment differently today than it once did, and it’s becoming more of a big risk. For people beyond a certain age, it may even mean permanent premature retirement. So, it’s really a pretty big looming risk in terms of the impact on consumption and so forth.
So yeah, I think awareness is building, and part of what I’m trying to do with this book is to begin a conversation. I think that that’s really important. One of the things that we’ve seen historically is that politics moves really slowly. I mean, I think you can point to healthcare as being a good example. Franklin Roosevelt was talking about universal healthcare in the 1930s, and it took something on the order of 80 years to get to Obamacare, which for the first time is some form of universal healthcare in the United States. Even now, it’s extraordinarily divisive; the Republicans would like to repeal it, some people hate it. But at least we have it now after 80 years, and that’s how long it took to get to the point where we were actually able to implement something whether you like it or not.
When it comes to this issue of technological unemployment from accelerating technology, I don’t think that we have 80 years to play with. I mean, this is going to happen a lot faster than that. And yet, our ability to adapt politically to this challenge is just extraordinarily slow–history shows that. So, I think it’s really important to begin talking about this, and to give it more visibility, and to hopefully kind of inject this question into the political arena sooner rather than later if we really want to have a good outcome to all of this.
Ted: Yeah, well we definitely agree with that. We’ve been pushing for people to talk more about this for some time. Before we wrap up, I want to give you a chance to respond to something–this is maybe a little bit of fun… But in that Hanson review of your book, Robin Hanson proposes two bets, and I want to read these bets and ask you if you’ll take these odds, because I think they’re interesting, I think they might lead us to a little bit of an interesting discussion.
So, he says, “I’ve expressed my skepticism about big automation progress soon,” and this is the bet that he has–he’s laid two bets: one is betting $1,200 at 12:1 odds that the BLS’s measurement of the labor fraction of US income won’t go below 40% by 2025, so below 40% in ten years. And then the other one is that he’s bet $1,000 at 20:1 odds that computers and electronics hardware won’t be over 5% of the US GDP–that is businesses purchasing computers and electronics hardware won’t be over 5% of US GDP in that same ten-year period. What do you think of those bets? And would you take those odds?
Martin: Well, yeah, I’d have to think about it. It’s possible. But in general, I’m not a huge fan of bets because of the famous story of Paul Ehrlich and the other guy, Julian Simon. So, Paul Ehrlich and Julian Simon had a bet, and it’s very famous, and Julian Simon actually won the bet. He bet against this thesis that we would see a rising commodity price as a result of overpopulation and stress on the environment. Conservatives look at that, and they believe that as a result of Julian Simon winning that bet, this issue has been put to rest. And yet people who have analyzed it found that if you looked at the outcome of that bet today as opposed to within the timeframe that they actually made the bet, it may have been the opposite result.
So, it really didn’t settle anything, because any bet is specific to a particular timeframe. So, for that reason, I’m a little bit skeptical to the idea of taking a bet with a specific date. It could be that by 2025, Robin Hanson is going to turn out to be correct, and maybe we won’t see that. But maybe by 2035, it’ll be a huge issue. I really don’t know exactly what the timing is going to be, but I do think that this is coming and that it’s inevitable.
Ted: Yeah, I think that’s a legitimate point. I also want to point out that I think this bet is extremely stacked. Like, it’s only a ten-year time period and he’s saying that the labor fraction of US income won’t go below 40%. Now, I looked this up and I couldn’t find numbers for this year, but the recent change over the last 50 years in that statistic went from 62% to 58%. So, there’s been 4% negative change, and he’s suggesting that it would drop an additional 18%.
Martin: Yeah, I mean if it falls to 41%, that would be pretty bad, I would think. I wouldn’t be happy if it were 41%, because I think that would be a pretty dramatic drop.
Ted: Right. Well, even if it were 50%, which would be I think pretty extreme compared to what the rate has been so far–like, I’m not at all convinced that 40% is a magic number under which we have economic unrest and over which we don’t. I think it seems to me like you could completely agree with your thesis, Martin, and still assume that 40% is a pretty extreme number for it to hit.
I don’t have good data on the other bet, but my guess is that it’s kind of moot, because he says, “Odds that computers and electronics hardware purchases…” He excludes software. So that, I think, is leaving out a huge question mark, which is that the hardware may become so incredibly cheap that even though we’re buying a lot of it, we’re not spending a lot on it, and we might be spending most of that money on software to make the hardware useful. So, I feel like he could easily win these bets without being right.
Martin: Yeah, when you read the bet to me I didn’t realize that was excluding software. Yeah, that would certainly be a big issue. I think that already software is becoming a much bigger component of IT spending.
But in general, yeah, Robin Hanson, I have kind of a history with him. I once actually debated him on a TV show about this, and he has kind of an odd perspective. He’s actually a Singularitarian, he believes in this stuff in the far future. He just doesn’t think that it’s going to have an impact now. So, he has I think an odd perspective, which says that eventually we’re going to have real artificial intelligence or everyone is going to upload their brain into the cloud or something, and then at that point it’s going to be a huge issue. But that’s not going to be for 100 years or something. But in the meantime, it’s not going to have any impact at all. That doesn’t make a lot of sense to me.
I think that long before we get to true artificial intelligence, or brain emulation, and things like that, specialized technology are going to really have a dramatic impact on the work that a lot of people are doing, and the reason is simply that most people do relatively routine specialized things. I mean, that was really Adam Smith’s whole point about the division of labor, is that over time people do more and more specialized activities, and the reality is that not everyone is a rocket scientist. A lot of people, a huge percentage of our population, do things that are, on some level, fundamentally routine and predictable. Displacing those people does not require science fiction artificial intelligence. It only requires machine learning and specialized artificial intelligence of the type that we see already and which is sure to get better and better over the next couple of decades.
Jon: So, I want to ask you one final question, which is given that all of us seem to support some version of a basic income, but given that this is not much of a often discussed topic in modern American politics–in fact, I think it’s the kind of thing that most people immediately have a negative response to, upon hearing the idea, if they’re not familiar with it… How do we go about selling this to people if it’s what we need moving forward? I mean, obviously one answer might be to write a book about it, which you’ve done. [Laughs] But beyond that, how do we persuade people that this is not some sort of socialist takeover?
Ted: Yeah, what’s the framing that we should use? Because I think, on the surface, it sounds so deeply un-American. But I think, as you’ve said, it’s potentially the only way to maintain our current competitive market social structure. So, how do we market that?
Martin: Well, one of the things that I try to do in my book when I talk about is I try my best to market it as a Conservative idea rather than a Socialist idea, and that’s actually the correct perception. I mean, the idea of a basic income has had support from people on both sides of the political spectrum, but actually it has been Libertarians and Conservatives that have really talked it up quite a bit in the past, and most notably was Friedrich Hayek, who has now become kind of a Conservative icon. You hear about Friedrich Hayek on Fox News all the time, but he was a big supporter of a guaranteed minimum income. Milton Friedman, another big Conservative icon, supported a negative income tax. Richard Nixon actually had an actual proposal for guaranteed income at one point.
So, it is an idea that can be embraced by Conservatives. The perspective I would offer is this: if this is really going to happen, and if we’re going to see people increasingly having their livelihood, their incomes impacted, and if people are going to become more and more insecure economically, eventually people are going to insist that the government do something.
There are two paths you can take: have the government take over more and more of the economy, create jobs for people artificially; maybe house people, build low-income housing, or even this kind of dystopian quasi institutional-type environment where people get housed if they don’t have jobs; maybe provide food and shelter to people, have the government do that directly. That’s really the big government/Socialism path.
The other path is to just give people money and let them go out and participate in the market. And that’s what Hayek supported, that’s what Friedman supported, that’s really the Conservative idea, is that you have a market-oriented approach. So, a guaranteed income or a basic income, whichever path you want to take, is really a Conservative idea when you look at it in that sense, and that’s the argument that I make in my book.
One thing that’s quite remarkable–an article I just saw a few days ago, it was on the site Vox, suggested that a couple of proposals from Republicans that are on the table right now, including one idea from Marco Rubio and also the FairTax proposal, which remarkably enough is supported by Ted Cruz I think, actually have embedded with them a path toward a guaranteed income perhaps. The FairTax actually–and I don’t know if that’s an idea that I support or not–but embedded in the FairTax is a proposal to give a rebate to everyone of about $7,000 a year, whether you work or not. So, that would actually be, in essence, a guaranteed income. I don’t know if the Republicans that support that idea are aware of that. Maybe not. And maybe if they were aware of it, they’d be unhappy.
Ted: Maybe we shouldn’t tell them. [Laughs]
Martin: Remarkably, there are a couple of Conservative proposals out there that could at least be some sort of a path toward a guaranteed income, whether the people who propose those ideas know it or not. So, there is more talk of this. There are proposals on the table that could lead in this direction, and I think that hopefully we will see more of a conversation about it.