Aug 13, 2018

Tim Lee on Technological Innovation, Productivity, and Economic Growth

Radical structural changes are coming to the transportation sector, and society should respond appropriately to these developments.
David Beckworth Senior Research Fellow , Tim Lee

Hosted by David Beckworth of the Mercatus Center, Macro Musings is a new podcast which pulls back the curtain on the important macroeconomic issues of the past, present, and future.

Tim Lee is a senior reporter for Ars Technica covering tech policy, blockchain technologies, and the future of transportation. Tim was formerly with Vox, where he covered tech issues and the economy. Tim joins the Macro Musings podcast to discuss technological innovation, economic growth, and implications for policy.

Read the full episode transcript:

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to macromusings@mercatus.gmu.edu.

David Beckworth: Tim, welcome to the show.

Tim Lee: Hey, thanks for having me on. I've been listening to the podcast since the beginning, so it's great to participate.

Beckworth: Well, it's fun to have you on. We have interacted before. You've actually written on some monetary policy issues as well as technology, so you have quite a unique skill set, and I think those two areas are going to be increasingly important, moving forward. And we will talk about this later. But before we get into all of that, why don't you tell the listeners a little bit about yourself. How did you get into technology, covering the economy, and just your general career path?

Lee: My background was in computer science. I got a computer science undergrad degree, but I was also very interested in policy and politics. And I spent a lot of time arguing with people about politics online in college. And so, then, after college, I had a friend who had a job at the Cato Institute. And I applied, and not really expecting to get the job, but got the job as a writer there. And so ever since then I've been looking for ways to kind of combine my interest in policy and technology. And, so, over the years I've spent some time in the think tank world, but recently I've been a journalist, focusing mostly on tech policy, but I also have done some writing about economic policy.

Beckworth: Yeah. So you're the guy to go to if you've got questions about the latest and greatest technical innovations, and that's why you're here. But you're also here, again, because you've thought about the implications for policy, which is an important issue. So, let's begin our conversation today and focus first on some of the recent technological innovations, what we've seen, what we know. And then, as we move on to the show, we'll tie it into some of the policy implications for both economic growth prospects as well as maybe monetary policy and other economic policy concerns.

Beckworth: But let's start, again, with the state of technological innovation. And you had a Vox piece that was titled Nine Radical Changes That Are Coming to Transportation. So, can you walk us through those and tell us why they are radical?

Lee: Yeah, absolutely. So, I think the interesting thing about transportation is there's been a long period now where, really, most transportation technologies have not changed that much. I mean, certainly there are new features to cars. You've got air bags, and antilock breaks, and stuff like that. But cars basically work the same way they did even 50 years ago. But that is now changing pretty quickly. The three big innovations that you see with cars... One is electric cars, led by Tesla, but a lot of other companies are also getting into that. Self-driving cars led by Google's Waymo unit. And then, also, ride-hailing, I think, is more about how people use cars, but I think is going to end up being very interconnected with that.

Lee: And so, in the last 10 years or so, we've seen all three of those go from really not existing at all, or being very marginal, to being very mainstream things, and I think most experts now expect, within the next 10 or 20 years, all three of those to be very important, if not dominant, ways that people use cars.

Beckworth: Okay. You also mention, though, flying drones. Tell us about those.

Lee: Yes. Absolutely. So, for passenger cars you have self-driving cars, but then, obviously, if you've got a self-driving car, you can make a smaller one that rolls on the ground and use it to deliver packages. And then you can also do the same thing with airplanes. You can have unmanned airplanes. And then, if you make those very small, you can deliver small packages. And so, I think it seems pretty likely, with a lot of startups now working on this, on a variety of robots, either rolling on the ground or flying through the air, designed to carry various kinds of cargo... Also, on a larger scale, people are working on self-driving trucks. So in the future, if you need to move a truckload of stuff across the country, you won't necessarily need a truck driver to do that. So that's another kind of category of major innovations.

Lee: And then we've seen some innovations just in air travel. There's a startup called Boom that's working on supersonic airplanes, which was a technology that had kind of an abortive step in the 1970s and '80s, with the Concord, but then proved not to be economically viable. And, of course, you've got SpaceX working on reusable rockets, which could open up the future for space travel, and potentially, in the very long-distance future... These are rockets that can take off, and then land, and then be reused. And Elon Musk likes to say that the current rocket industry, before SpaceX, was like if you had to build a new 747 every time you wanted to fly from New York to LA. Once those rockets are reusable, the fixed costs, the capital costs, of space flight become much cheaper, and so you could imagine space tourism.

Lee: And you could also just imagine a future where, if you have to go from New York to Hong Kong, you maybe have a quick boat ride out to an off-shore launchpad, and it might take half an hour or an hour to go to Hong Kong. There's a lot of practical reasons why I think that's pretty far away. But that's an example of something, maybe in a 50-year time horizon, if reusable rockets become very practical technology, you can imagine people experimenting with.

Beckworth: Okay. And the final one, I believe, in your list was the Hyperloop. You mentioned trucks already. Now, the nice thing about this list is transportation. It's technical innovation in the physical world. And later in the show I'll come back and discuss some of Michael Mandel's work on this. He kind of argues that we're really waiting for the innovation to really translate into the physical processes. And this'd be a big part of it, right? Transportation as a physical activity. So let's walk through this list, some parts of it, and talk about how it will change our lives.

Beckworth: And let's start with self-driving cars because that's probably what everyone's most familiar with, our listeners, and it's kind of been a hot thing. So, you had a article titled Self-Driving Technology's Going to Change a Lot More Than Cars. So, let's start with cars, though. How will our lives change in a meaningful way because of this technology?

Lee: Well, I think the most obvious and direct result will be that if you have a morning commute, you won't have to pay attention to the road, and so you'll be able to read, or watch TV, or do your hair, or whatever. So that'll be some time people get back, and also, hopefully, the crash rate will go down a lot, and so we'll save fewer lives. But I think potentially the larger implications will be for changing how people use cars and how we sort of organize cities.

Lee: I mean, if you think about the invention of the car originally... People obviously thought a car was a major invention. I don't think people necessarily anticipated the invention of suburbs, the emergence of stores likes Walmart. There are a lot of kind of second order consequences. And so, one example, for driverless cars is it's likely that parking lots will become much less common because a lot of people are likely to hail cars on demand rather than owning one. Because without a driver those will be much cheaper. And then, even if you do own the car, it can kind of drive to a parking space, and so you just won't need as many spaces.

Lee: And so this could solve a number of problems. I mean, certainly in urban areas, one of the big objections to denser development is, "Well, what about street parking?" But obviously, if there's a lot less demand for parking, that could allow for denser development, and it could just allow different development patterns. Right now you have high density just near subway lines or bus lines, but you could allow different kind of development patterns in other places, where you don't necessarily need a car in order to get around.

Lee: So that's on the car side. Then, I think it's going to have really big impacts on the retail side. Because right now if you want to buy something with e-commerce, you go to Amazon. It probably takes two days to get your package. But if you have these autonomous delivery drones, it could easily take an hour, half an hour, to get your package. I think once that becomes commonplace, it really won't make sense to drive to a local store, in a lot of cases, if you can instead have the store send the goods to you. And I think it's hard to say exactly what the structure of that industry will be, but I think that the big-box store as we know it is likely to radically change or maybe go away altogether.

Beckworth: Huh. So, the Walmart shopping experience may become a thing of the past.

Lee: Yeah. Absolutely. So, if you think about the Apple store, that's really more kind of a customer service... There's a lot of training. So I think some kinds of stores will probably be fine, but yeah. A store like Walmart, that is... that you're not going there for the great customer service or the ambience... I think what might happen, actually, is those might gradually turn into warehouses for the self-driving drones. The drones won't have very long range. You'll still need warehouses, and so you'll have these big boxes. They could be perfectly good warehouses. But it just might be a lot fewer people actually go to the store. Because why? What's the point if instead you can have the company do the work, sending the product to you?

Beckworth: Sure. Yeah, I thought maybe Walmart days were numbered, or retail in general were numbered, with the advent of 3D printing at some point. A lot of the things you might buy from them, you might be able to produce at a local 3D shop or maybe even at your house. But this car angle is a unique one I hadn't thought about. So, the driverless cars, drones... They themselves would become the form of transportation, where you wouldn't actually have to take time out of your day to go do it. So that would be a radical change. I want to go back to the garage. There'd be fewer garages in downtown. But even homes, like homes in the suburbs or rural places... You wouldn't need to build garages anymore, right?

Lee: Yeah, potentially. I think rural areas it'll probably take longer because... I mean, if you go to kind of a ride-hailing model in a very rural area, there just won't be a lot of cars around, and so it probably will make sense to still own a car. But yeah, so, I think more and more new houses, certainly, it just might not make sense to devote the money to having a driveway and a garage if there's just a kind of fleet of driverless cars available for you to take whenever you need one.

Beckworth: Yeah. Couple other questions about this revolution in cars, transportation. So, one of the issues I see, and I've heard others talk about, is just kind of the culture behind accepting it. I heard another show talk about how elevators used to have operators. To get in the elevator, when it first came out, was real novel, almost scary, experience. You had to have someone in there to press the buttons for you to go up and down. And it took some time before anyone would actually have the trust, or the belief, it was okay to get an elevator without an operator.

Beckworth: And this show compared that to airplanes, and they made this argument... And I'm not an expert to really verify this, but their argument was, in theory, you could have pilotless airplanes right now, but we never got past that fear, that hurdle. And so the question is, will cars get to the point where they're more like the model that the elevators followed, where eventually we got past our fear and became a culture where that was accepted? Or we'll be hung up on safety. "We've got to maintain control. We're the pilot." Where do you see that coming out?

Lee: I think there definitely will be a transition process, where you'll see a mix of vehicles, and I think probably older people will be more reluctant. Obviously, a problem that many of us will have as our parents get older is, at some point, they might not have the eyesight or the reflexes to drive. And so, then, if you're facing a choice between not having a car at all or being able to take a driverless car, I think many people might accept it. But I think the difference with airplanes is with an airplane... Obviously, pilots make a lot of money. But, even still, the pilot is not that large a fraction of the cost of the airplane.

Lee: And so the potential savings, I think, are maybe just not worth the kind of bureaucratic hassle and the PR issues, whereas with a self-driving car... If you look at Uber, a taxi ride... The driver's pay is about half the cost of the ride. And so I think if you can offer taxi rides at half the going rate, a lot of people'll be able to give it a try. And then, over time, I think, as it becomes more widely used, it'll naturally become more accepted, and certainly if data starts coming out that it's safer, which I don't think... This definitely hasn't been proven yet, but I think, 10 and 20 years, that's likely to be the case. Then, I think, that could even work in the other direction, where there's a lot of pressure for people to stop driving because it's actually not as safe a thing to do compared with the alternative.

Beckworth: Yeah. Now, I could definitely benefit from a driverless car. A lot of work I could get done in the vehicle when I'm driving to the airport or driving to downtown. I would definitely benefit from it. I would be eager to embrace it. I'm trying to think. The places where I would have reservations might be if I send my kids to soccer practice in a driverless car. Their first few times... They're in the car by themselves, let's say, and they go off. I might be a little reluctant. But I guess one can get used to that.

Beckworth: One other area, though. You mentioned earlier there might be some holdout. Again, I think of rural areas, where there is maybe this spirit of freedom, this cowboy culture. I hate to use that term, but kind of like, "I like my car." I'm thinking of someone in a pickup, but even someone that owns a sports car, right, who wants to just be able to hug a corner, go real fast. How will those holdouts integrate into a world where, if you drive, in most cities you've got to be a part of some kind of network or some kind of smart machine? Will we ever find a way to reconcile those two views?

Lee: Well, so, I think one of the things to keep in mind about self-driving cars, especially in the early days, is that, obviously, the first one will have to share the road with a lot of other cars, and pedestrians, and so forth and is going to have to be able to operate without everybody else adopting any kind of coordinating technology.

Lee: And so I think that'll just continue to be true, that self-driving cars will just know how to deal with human drivers. And the other thing, I think, to keep in mind here is that with any technology, it takes, typically, 20, 30 years to go from a totally new technology. You think about cellphones were new in maybe the '80s, and they only became totally ubiquitous maybe five or 10 years ago. So, even in the best case, even assuming everything goes the way advocates hope, it'll be well into the 2030s, 2040s before it's kind of the dominant mode of driving.

Lee: And I think there will probably be a generation gap, where people who are in their 50s, 60s, 70s now... Some of them will just never be willing to do a self-driving car, and that's fine. They can keep driving their old-fashioned cars. But people like my kids, who are very young right now... They will grow up in a world where cars have just always driven themselves. My parents read a newspaper. They'll probably read newspapers till they die. And that's fine. Sometimes it just takes a little bit time for attitudes to change as a new cohort comes in.

Beckworth: Yeah. There's a lot of things even my kids today don't know that I did in the past. VHS tapes. Things that would be completely strange and foreign to them. Cassette tapes. But, okay, so, it takes time, technology, to be implemented and widely diffused throughout the economy. Let's look at some of the other examples and ask how they would also affect our lives. So, tell us about the Hyperloop. How would that make us better off or change the way we live?

Lee: I would say this is the one that I'm most skeptical of, from a technical perspective. So, the idea of the Hyperloop is this sort of a super-fast bullet train that works by running the train through an evacuated tube. And so, because there's no air resistance and it's magnetically levitated, it can go very fast. It's a technically elegant solution, but I think the difficulty is that even with existing train technology... If you look at the Acela corridor between Washington and Boston, the reason it doesn't go as fast as the trains in Europe is mostly that the tracks are not straight. Because the kind of political will to get the straight right-of-ways you need was not there. And same is true of the Hyperloop. You can't go fast unless you can go straight.

Lee: And, theoretically, because the Hyperloop tubes are a little narrower, it might be easier to get right-of-way because you don't need it to be as wide. There's a number of people working on this. I would love to be proven wrong, but I think it's... Conceptually, it's a faster version of a train, and so, if it worked, you could imagine people commuting from Los Angeles to San Francisco, or from Houston to Dallas, that kind of thing, on a daily basis. So it would kind of knit the world together more than it is now.

Beckworth: Well, that'd be amazing, if that could happen. What about flying cars or autonomous flying vehicles? You had an article, recent piece, titled Electric Self-Flying Vertical Takeoff Taxis Are Coming to New Zealand. So, something else that comes to mind is we're thinking, I think, right now about using the roads and kind of on the ground, but there's a whole space above us, right, that could be used, if done properly, to facilitate transportation.

Lee: Yeah. So, I would say this is the confluence of a couple of different trends. Oh, I mean electrification and self-driving, those two technologies together. What electrification really does is, instead of having one noisy, giant gasoline engine, like a traditional helicopter... That is very expensive. It's noisy. And it's failure-prone. With an electric drone, you can have maybe 10 or 12 rotors. Even if a couple of them fail, it can still fly. So if we had better batteries, we can do that. But then it's very hard for a human being to control 10 or 12 rotors, but if you have software flying it anyway, software can very easily do that.

Lee: So you add those all up, and you have the potential for very small one-, two-, three-, four-passenger flying vehicles, with a range in the tens of miles and that's able to take off and land vertically. And so you could imagine a world where, the downtowns in cities, on the tops of the skyscrapers, there's what Uber calls vertiports, kind of like tiny airports, and people take off and land. Helipads, basically. And then they can go to a parking lot in their suburban town, for example. And so, rather than it taking half an hour, or and hour, to come in from the suburbs, maybe it takes 10 minutes. The technology is definitely not there yet. But there doesn't seem to be any technological barrier other than the batteries, I think, would have to be a little better for this to be really practical. But batteries have been steadily getting better over the years.

Lee: And then there's a lot of regulatory questions about are people going to want all these things flying overhead? Can they do it safely? And can you rearrange the airspace? Because right now the airspace assumes that every vehicle is manned, and it's really designed for a much lower volume. This is also an issue with the drones, with the delivery drones.

Beckworth: Sure.

Lee: The way the FAA regulates the airspace right now, it assumes a relatively small number of large, human-piloted drones that fill out paperwork, and have a flight plan, and all that. And so, to have a bunch of small autonomous vehicles, there's going to have to be a new infrastructure, and a new set of rules, about who gets to use what altitudes and how do we make sure they don't run into each other. I think that's going to be the big barrier, and this is a big part of the reason why they are testing this stuff in New Zealand. Another one of these flying drone companies is testing in Switzerland. So I think some other countries have just more flexible regulatory regimes versus what the FAA does here, and so they're testing it. But if it becomes very successful overseas, obviously, there will eventually be a lot of pressure to liberalize here and allow similar kind of technologies here.

Beckworth: So, talk us through what's going on in New Zealand and Switzerland. These are vehicles that you would literally get into. Is it automated? Is this machine just... lift you off in the air? So someone in New Zealand, as we speak, is taking a vehicle through the air, putting full trust in machine?

Lee: I'm not sure if they've actually launched, but the basic idea of the New Zealand test, I think, would be, yes, you would hail a vehicle. I'm not sure if it would go to your house, but you'd go somewhere nearby.

Beckworth: Sure.

Lee: You would get in. And I'm not sure how many passengers, but the idea is, in the long run, there would be no pilot. And so, potentially, you'd just get in all by yourself, and it would just fly itself up in the air, fly to the destination, and land. And you would have an app, where you say, "Here's where I'm trying to go," and it tells you which vehicle to get into.

Lee: The situation in Switzerland is different. That's a company called Matternet that's doing deliveries. I believe they're ferrying blood samples between hospitals. And so right now the hospital will pay a courtier... I don't know... 20 bucks or something to ride a bicycle between hospitals, or drive a car between hospitals, to deliver blood samples. And they can do that for a fraction of the price, and much faster, by flying from the top of one hospital to the top of another hospital. And this seems like a model that can scale up very, very quickly. And you could end up first delivering high-value cargo like medical equipment, but eventually just your Amazon package could just be dropped on your roof or in your backyard instead of being carried by a UPS truck.

Beckworth: Yeah. These are all very fascinating and exciting developments. And, again, it'll be neat to see how our children live differently than us, when these things have all come to manifestation or to fruition. Let's talk about some other areas of technological innovation. You focus more on transportation, but some of the other ones that people often talk about... I mentioned one already... 3D printing, robotics, artificial intelligence. With medicine, there's CRISPRs, gene editing, which seems very promising. Maybe even renewable energy. I mean, do you see a lot of progress in those areas and if not progress, at least great potential for them?

Lee: I would say it varies. So, 3D printing, I don't think that there's the huge potential that people thought. And I think this is becoming the conventional wisdom. If you think about when regular, 2D printers came out, you might've thought, "Well, in the future nobody'll print books in old-fashioned factories because we'll just print out our books." But it actually turned out it's a pain, and expensive, to print out your own books. And at a certain scale it's actually more efficient to use traditional mass manufacturing.

Lee: And I think the same is true for 3 printing. For consumer uses, if you want a plastic toy, yes, you could buy a thousand-dollar 3D printer and some plastic, and download some CAD file, and whatever. It's pretty hard, actually, to compete with Amazon, with their free shipping. Where 3D printing is proving very useful is in industrial applications. If you're doing prototyping, it's really useful, if you have some part you're building, to be able to bang out a copy of it and look at it. It's an important technology that I think will become more important over time, but I don't think it's going to have the kind of global economy scale impact where traditional manufacturing becomes obsolete.

Lee: Robotics, I actually feel like is something that people talk about a lot, but it's actually pretty old technology. The car industry, for example, has had robotics since the '80s or '90s. It's definitely continuing to be more sophisticated. As the software gets better, the robots are getting better. But, at least in some areas, like in auto manufacturing, I feel like some of the low-hanging fruit was plucked a long time ago. I mean, you actually see this with Tesla, which... A couple years ago, Elon Musk was bragging about how they're going to have the most automated factory ever, and they're going to have this lights-out factory. And that basically turned out not to be true. The last few months he's admitted that actually, "We did too much robots." And they are still, obviously, doing a lot of automation, but this is not something where I think you're about to see a quantum leap. It's just continued, incremental progress of the type you've seen in the last few decades.

Lee: Artificial intelligence is definitely... I mean, I guess that's a broad category. So, self-driving cars is one example. Definitely, image recognition, speech recognition, you've seen big leaps already. I expect continued improvement. And so I think there are lots of areas. Five or 10 years ago, when they invented the smart phone, there were a bunch of apps that came out that enabled things like Uber and Lyft. I think artificial intelligence will be a category like that, where there are lots of things that are apps that maybe come, voice-activated type things or lots of business processes that become more efficient, when you can have a camera that recognizes if a customer comes in the door or whatever. But I think it'll make our lives better in lots of small ways, some of which we might not even notice and some of which we will.

Lee: CRSPR definitely seems important. I don't know very much about medicine, so I don't have a specific opinion about it, but that definitely seems... I would say, the big picture, the two major categories of innovation that I think are significant are transportation and medicine. And if you think about the problems people have that technology can solve but are not solving right now, one is we'd like to get from point A to point B faster, and the other is we'd like to live longer and get sick less often. It's actually hard to think of that many other ways. If you're in the upper-middle class, you have pretty much all the food you want. You have pretty much all the clothes you want. You'd like a bigger house, but mostly people want to live in better neighborhoods rather than necessarily having bigger houses. But I feel like transportation and medicine are the major areas where there's likely to be significant technological progress.

Beckworth: That makes a lot of sense. I mean, firms are going to invest in areas where they get the biggest return on their investment, and transportation and medicine seem to be an area ripe for further advances in technology. But let me ask-

Lee: Oh, you-

Beckworth: Go ahead.

Lee: You mentioned renewable energy. I do think that's another area. So, we talked somewhat about batteries. Batteries and solar power are both... I mean... I think a similar place, which is that they've been around, obviously, for many decades. But there is a basic kind of economic process where, in positive feedback, we can scale on price, where as you produce something at higher scales, you get economies of scale, and so the price goes down, and that then expands the market, and the price goes...

Lee: So, with batteries, cellphones really jump-started demand for batteries, which then allowed them to be more affordable, and that in turn made them... more economical to have electric cars, which use massively more energy than cellphones do. And as they scale that up, batteries will get cheaper still, which will then make it feasible to do for, for example, the electric grid. It'd be very useful to have batteries to back up solar panels, and that's not really quite economically feasible yet, but in five or 10 years it might be thanks to the increased scale from cars.

Lee: I think you see the same kind of thing with solar panels. You have this virtual circle where it gets cheaper, and so more people use it. In some of your places like Arizona, it's pretty close to the point where, even without subsidies, it's price competitive with conventional utilities. And, a few more turns of that kind of virtual circle, you could end up with a case where it's just straight-up cheaper in many parts of the world.

Beckworth: Okay. So, one area where we haven't seen a lot of change, and a lot productivity growth, is in the building of homes. I know part of the steep price of housing often, in cities, is zoning restriction, NIMBYism, regulatory issues more than anything. The price of the land itself. At least, it seems to me that homes are still built fairly similarly to what they were before. You could imagine, though, at some point, robots coming in or some kind of automated process where homes are built quickly, cheaply. Is there any prospects along those lines?

Lee: I'm not expecting big changes there. So, one thing to think about there is that usually automation is most effective when it's done in a controlled environment. In a factory, ideally, or in a situation... Like, trains go on tracks, which allow them to kind of work better. And homes are inherently a non-standardized situation. You're dealing with large-scale materials, where a lot of the work is transporting materials to the location and moving large things around. And so it seems like it's not likely to be very easy to build robots that can do on-site manufacturing.

Lee: Now, of course, we do have manufactured homes, mobile homes, which I think it's mostly for regulatory reasons that those... Oh, I mean, partly people don't necessarily like them, but also I think there are a lot of regulatory restriction, where I think you could have a lot of affordable housing if you let people set up more trailer parks, and live more densely in trailer parks, but I think there's a lot of legal restrictions on doing that. If you had pre-fab houses, I think that would probably be seen as downscale, and you'd see regulation to restrict it.

Beckworth: Yeah. So, the bigger issue there is, I think, regulatory issues more than anything else.

Lee: Absolutely.

Beckworth: Okay. So, we've touched on technology. It seems promising. The future looks bright for our children, absent some horrible war or something else. It seems like there's a bright future. But, as you know, Tim, if you look at the data, it's really hard to find any evidence of the technological progress to date that we've talked about. If you look at total factory productivity, which is probably the most carefully thought-about measure of productivity, it's been in a slowdown since the 1970s, with a few exceptions. The late '90s, early 2000s, there's a bit of a blip. But we just haven't had the rapid productivity gains one would expect to have seen given some of this... technological advances that we've talked about.

Beckworth: Peter Thiel has this famous saying, "We wanted flying cars. Instead we got 140 characters," talking about Twitter. There's a number of naysayers out there. Robert Gordon has this... It's a good book, but it's very pessimistic about growth. He says in the book, "We don't eat computers, or wear them, or drive to work in them, or let them cut our hair." I really think he's way too pessimistic and misses some of the issues. But there's a lot of other folks who share this view. Initially, kind of the pushback that I bought into is it's just being mismeasured. There's been some people who have argued, "It's not a mismeasurement problem." Chicago columnist Chad Syverson, and then San Francisco Fed study also came out with this. Then, finally, my boss, Tyler Cowen, has kind of made this argument several places, and most recently in his book, The Complacency Class, that our nation is just becoming increasingly risk-averse. We've been successful. We've had some material affluence, and we're just risk-averse in everything we do, from raising our kids to how we build things to how much progress we can tolerate.

Beckworth: Now, those are the naysayers. There seem to be a lot of them, and the data seems to be consistent with their view, but there's also the true believers. I mentioned, earlier, Michael Mandel and Bret Swanson. I had them on the show earlier. They had a paper, The Coming Productivity Boom: Transforming the Physical World With Information. And they basically argue that there's been this information gap, that a lot of the gains have been limited to idea space, software, but eventually... And we've touched on some of these. The transportation. They'll be seen more and more in the physical world. When that happens, watch out.

Beckworth: So I want to hear your take on this. And you had a great article you wrote, The Productivity Paradox: Why We're Getting More Innovation But Less Growth. But, just in general, what is your take on this concern that we don't see evidence of it in productivity growth, and will we ever?

Lee: So, I would say overall I'm in the Robert Gordon camp in the sense that I feel like it's really useful to think about the big picture. Say you're looking at 1900, and you think about what are the ways that technological growth, material progress, could improve people's lives. A lot of people didn't have flush toilets. A lot of people didn't have enough food to eat. A lot of people had very small houses. Refrigerators weren't invented. Airplanes weren't invented, et cetera. And so it's very easy to imagine, in an optimistic future, how could people's lives get better by giving people more stuff, more technology.

Lee: And, as I was saying before, if you think about that same exercise now, a lot of parts of our lives... I mean, they're not perfect, obviously, but the problems we have are mostly not problems that technology can solve, with the big exceptions, I think, of transportation, medicine. You could probably think of two or three others. But if you think about somebody who, say, makes $100,000 a year and gets a big raise, what do they do with the extra money? They probably move to a better neighborhood. They maybe put their kids in a private school. If they're really rich, they maybe buy some rare art. There are a few things. Rich people, I guess, have private planes. But there are really not that many material things that wealthy people have that middle class or poor people don't have.

Lee: From a big-picture perspective, I think Robert Gordon's correct that there was a series of just transformational inventions in the late 19th and early 20th century that moved us up to a pretty good place. And it's just hard to imagine what the subsequent things will be. Now, on specific things, I think, he was too pessimistic. I mean, self-driving cars... I feel like he did not have enough of a sense of kind of imagination of how that could... the second order of facts of that. Obviously, people like to have longer lifespans, and there are diseases like Alzheimer's that we'd love to cure.

Lee: So, at a high level, I think it is not surprising that the rate of technological growth has slowed down because I think the remaining problems humanity have is not necessarily that type. I would also say that I think there's a kind of self-limiting character to technological... That was the main argument of the article that you mentioned. You think about something like food or clothing. Those are both, obviously, very important. And there has, in fact, been some continued progress in the sense that farmers and textile manufacturing has gotten more efficient. But, at some point, what happened is it's just such a small portion of your budget, the average household budget, that even if you, say, doubled the output, it would have a very small, marginal impact on people's standard of living.

Lee: And the things that loom large in people's budgets are housing, education, and those are things where the main barriers aren't necessarily technological. They're either regulatory, or just you're buying a positional good. There's only so many spots in Harvard, and so making another school isn't going to work because that school's not going to be Harvard.

Beckworth: Yeah. So, you're saying that there is some kind of limit, then, to how much technological innovation there can be for humanity or practically.

Lee: Yeah, well, I guess I would say I think there's a limit to how big of an economic impact it can have.

Beckworth: Okay.

Lee:  You think about it also with computers. Actually, Moore's Law is slowing down a little bit, but Moore's Law went for 50 years. So computers are a million times faster than they were 20 or 30 years ago. We're not a million times better off because the marginal value of one extra compute cycle is much lower than it used to be. And I think that's just inherent in most types of production. The marginal returns go down.

Lee: Another concept that's really important here, I think, is Baumol's cost disease, which is this idea that... I guess it's the same idea, basically... that as material goods get more affordable, the services component of the goods and service you buy becomes a larger and larger fraction, and so continued progress in manufacturing is going to move the needle less and less because it ends up being just a trivial part of the overall cost of whatever you're producing.

Beckworth: Okay. Tim, it's interesting you would bring up this argument because I had a guest on early in the show, and this show was about Star Wars, and Star Trek, and macroeconomics of those shows, kind of a fun diversion.

Lee: Yeah, that was great.

Beckworth: Actually, I had two guests on. The one gentleman, who was talking about Star Wars, argued that secular stagnation had kicked in the Star Wars universe, that if you look at all the movies, you don't see a lot of major change technically. And then, he argued, therefore you're not seeing rapid economic growth as well. The argument you just made... I'm not sure if it's the same argument, but it seems to be echoing a similar point that at some point there's a limit to how much technical innovation can change the rate of economic growth. Is that right?

Lee: Yeah. Yes.

Beckworth: Okay.

Lee: The rate of economic growth. I think that I, personally, the things that I could... the way that I could benefit if I had more money... They're mostly not that I want more food, or more clothes, or another car, or something like that. It's mostly sending my kids to a better school or living in a nicer neighborhood. Any individual person can buy those, but it's not necessarily something that everybody can have.

Beckworth: So, measured productivity may not change a lot. Consumer welfare might go up. Well, let me ask you this question. Let me push back just a bit. Can you envision anything that would really push up economic growth? Any kind of technical innovation, no matter how farfetched it was. Can you think of a scenario where this would be the case?

Lee: Well, I think, certainly in the short to medium term, some of these transportation innovations could. So I think that if driverless cars are successful, I think we very well might have another decade like the 1990s, where a bunch of industries are being transformed at the same time and certain types of goods get a lot cheaper. And then, obviously, if we get practical space travel, that will open up a whole new area that could potentially have an economic boom. So I'm definitely not of the opinion that there will never be rapid growth again. If you think over the next century, my guess... And obviously it's hard to predict, but my guess is that the overall rate of growth is not going to be anything like it was in the 20th century, for basically the reasons that Robert Gordon articulated.

Beckworth: Okay. Well, let's move on, then, from that, and let's talk about the policy implications of technological change. Because one of the things we do know... even if there's not going to be rapid economic growth... that technological change can be very disruptive. If you're in the industry that is dislocated, if you lose your job because a robot takes over, artificial intelligence takes over... If you were a truck driver, and now there's these driverless trucks, for example... And truck driving is a very rewarding profession currently, but that could be automated. So how should society handle such disruptions? I mean, people talk about, in extreme case, if great singularity comes. But how would you see policy responding to the structural unemployment that would be created by these changes we've talked about?

Lee: So, one thing I think is interesting to talk about here is I think we're starting to see some sign of how, for example, the self-driving car industry is going to be structured. And I think there's some sense that the short-term job losses are not going to be as significant as some people feared. And you've seen some research about this in the past. So, for example, Jim Besson, an economist, has made the point that when the ATMs were created, you actually didn't see a decline in bank tellers, and the reason for that was that because it became cheaper to open a bank branch, banks opened more branches. And so, even though there were fewer tellers per bank, there were more banks, and so employment went up.

Lee: I think in the case of self-driving cars, especially because the initial business model is likely to be ride-hailing, where it's a taxi service like Uber rather than something you purchase, there's going to be a lot of associated jobs, even though there's not going to be somebody in the car, driving the car. There's going to be people who are going to need to provide the customer service. If there's a problem, you need to call the company to say, "My self-driving car isn't working." You're going to need mechanics. You're going to need people managing the fleets, and obviously you're going to have engineering jobs, jobs mapping the environment.

Lee: And so, I think, in the short term, and even the medium term, I think a lot of these innovations might actually lead to increased jobs. And in the case of trucking, there's actually a pretty severe trucking shortage right now, or truck driver shortage. In the longer term, definitely there's going to be some disruption. And here, I guess... to kind of go to the core Macro Musings topic... I really feel like monetary policy and macro issues are not discussed enough when talking about these issues.

Lee: So, if you think about it from a very high-level perspective, if you have a new technological innovation that increases productivity, society has a choice between making the same amount of stuff with fewer people or making more stuff with the same amount of people. At a basic level, you can think of monetary policy as a knob you can turn, choosing between those two alternatives. And if your fear is that you're going to have a future where there's such rapid progress that there's not enough jobs for everybody, that's just another way of saying that there's not enough demand in the economy. And we know how to create more demand, assuming there's not liquidity trap or something. But in normal times, we know how to do that. And sometimes I feel like if you're having technological unemployment, that's just another way of saying the central bank is not providing enough demand.

Beckworth: It's interesting you bring that up, Tim, because we are kind of going through an experiment right now with President Donald Trump's big budget deficits. And what will they do to the economy? And just last quarter we found out that GDP grew at 4.1 percent, which is a really great number. The question is, "Will these policies actually increase the productive capacity of the economy or not? Is it just a temporary, one-time increase?" We've often, on this show, talked about negative hysteresis, where you have insufficient demand, persistently, that it leads to an eroding of the capacity of the economy. But can you go the other direction? And what you're saying is, "There's some limit to this. But if demand was run a little bit hotter, would it build up productive capacity of the economy?" Is that right?

Lee: Yeah, exactly. And I feel like the last decade has made me much more favorable. Five or 10 years ago I would have been more the traditional economic view that demand side policies can't really increase supply outside of a pretty severe recession. But we've had this experience for the last 10 years, where people keep saying, "Oh, we're close to full employment," and then employment keeps growing, and we still have this slow, sluggish growth. And it makes me wonder if the Fed had pushed a little harder, two, three, four, five years ago, if we wouldn't have reached the same, current level of employment and output a little earlier.

Lee: And then I think it's actually pretty easy to tell stories about the mechanisms, how this will work. So, if you think about if you've got a factory, and you have an idea for an innovation that's kind of risky, but if you do it you could double the output of your factory per worker. But there's not a lot of demand out there, so that would mean maybe you have to lay off half your workers, or you would have to cut your prices to start a price war, or whatever. And it's like in a low demand environment, even though it might be good for your shareholders, as a manager it might not be that appealing to do this. And so a lot of companies might just be risk-averse and say, "Well, I'm just not going to take that risk." Whereas in an environment where it's hard to keep workers and demand seems to be growing robustly, it just is going to make a lot more sense to invest in expanded output.

Lee: And I think you've seen some of this in the news recently. There have been a lot of stories about companies willing to hire people they wouldn't have hired in the past, that maybe I don't have a high school or college degree, or have a criminal record. Companies more willing to do training. I think the substantive behavior of companies change in a tight labor market, in an environment with a lot of demand, compared to the kind of economy we had in 2010, 2011, 2012.

Beckworth: Yeah, I think another great data point that supports that view is the number of people on disability has been falling as the economy has been heating up, right? So these people would be considered structurally unemployable, but here they are finding jobs, either through their own effort or because firms are accommodating them in some manner, and so they're actually entering the labor force, which is a positive thing. So-

Lee: Yeah.

Beckworth: A technical economist way of saying this is that potential GDP is endogenous, at least to some extent. I mean, you can go to extremes and have 19070s inflation, but short of that, what you're arguing is, "Look. There's some true ceiling that we haven't hit." I've had some previous guests on who've made this argument as well.

Beckworth: And kind of the best analogy I can think to describe this is... And, Tim, I'll use you as an example, okay? Let's say, Tim, that you used to bench press 500 pounds. And you were just really ripped and strong. You could bench press that 500 pounds. You go to the gym, setting all kinds of records. And then you got sick. You were out for six months. You lost a lot of weight. And you go back to the gym, and you can only bench press 200 pounds. Right? Your potential is, at that point, 200 pounds. Before, your potential was 500 pounds. And if you settle at the 200 pounds, if you only work out enough to maintain that 200 pounds, you don't run yourself hot, you don't try to push yourself harder than the 200-pound limit, you'll always be stuck there.

Beckworth: But if you go back and say, "Look. I was 500 pounds at some point in my bench press. I want to get back there. I'm going to have to work extra hard. I'm going to push myself beyond my breaking point," then you might get back up to that potential. So I think that's kind of the best analogy I have in mind for this. And, again, this is, I think, a great period, where we'll be checking this out. We'll see if large budget deficits... if the Fed tolerates a little overheating, if it will actually happen.

Lee: Yeah. One other thing I'll say about this. I had a piece kind of arguing that there's limits on how much growth you can get. Because people just spend more and more of their money on things that are not manufactured goods. But if that were the whole story, one thing that you should still see is... You could still see rapid manufacturing productivity growth because the economy still could be producing more and more stuff, and people at the lower end of the income spectrum... They might not be able to afford to send their kids to Harvard, but they should at least be able to see more and more food, and clothing, and stuff.

Lee: But if you look at the data, at least last time I checked... It was few months ago... you're not seeing rapid productivity growth even for manufactured goods, which suggests to me that there is something going on beyond just a basic Baumol's cost disease kind of problem. And it makes me think that if you had more demand, that there are still lots of people around who would like more clothes. Not, I think, people in the upper, above-average, income, but people in the lower end of the income spectrum definitely could use more material goods. And the fact that it really does seem like incomes have stagnated for a lot of people, and they don't feel like they have that abundance of even kind of basic material goods, suggests to me that with more demand you should see more manufacturing growth, more productivity growth.

Beckworth: Yeah. Let me throw a different explanation out for why we haven't seen rapid productivity growth in numbers, and this kind of goes back to... Well, it complements Tyler Cowen's risk aversion story. Even if we could get rid of risk aversion in our society, maybe there'd be a limit to how much added productivity growth it would bring. But Tyler describes a nation where we've become just more and more risk tolerant. We're all becoming helicopter parents. We don't tolerate losses on job sites. We could never rebuild the Hoover Dam if we had to.

Beckworth: And, in that vein, you have a article on Vox where it's titled The End of the Internet Startup. The subtitle is We Haven't Had a Major New Technology Company in More Than 10 Years. And let me just read a few paragraphs and a few excerpts from your piece. You say, "Silicone Valley is supposed to be a place where a couple of guys in a garage or a dorm room can start companies that change the world. It happened with Apple and Microsoft in the 1970s; America Online in the 1980s; Amazon, Yahoo, and Google in the 1990s; and at Facebook in the 2000s. But in the 2010s, we seem to be suffering from a startup drought. People are still starting startups, of course, but the last really big tech startup success, Facebook, is 13 years old."

Beckworth: And what you go on to do is suggest an explanation that these established firms are taking measures to kind of keep out competitions. They buy up rivals. They buy up competitors. They try to maintain their market share, which, again, speaks to this, maybe, risk aversion, more rent seeking going on. Do you see that as a big reason for the slow productivity growth or just maybe a small part of it?

Lee: I would say it's a relatively small part of it. I mean, I think that that argument is mostly specific to the kind of Internet economy. So, I think you are seeing some startups... in driverless cars, for example... I do expect significant growth there. So, I think that that story's partly about just kind of the maturing of the Internet economy. With any industry, there's going to be a period of 20, 30, 40 years where there's a bunch of new companies that go on to become big companies, but then, at some point, the companies become mature, and those big companies then do various things to maintain their dominance. I do think there's a larger question about the kind of increasing consolidation of companies across a number of industries, and I have not made up my mind what I think about this. So, there's a certainly a school of thought that says a lot of different industries are increasingly centralized, globalized, and that that acts as a barrier to innovation.

Lee: As I think about the Internet economy, I personally think it was unfortunate that Google's acquisition of YouTube and Facebook's acquisition of Instagram and WhatsApp, Amazon's acquisition of Twitch... There's been a bunch of acquisitions like this. It's just interesting to imagine an alternative future where those were all independent companies. Regulators could have said, "No, you can't buy those companies." And I think most, if not all, of them would have gone on to be very successful. And in that world, I think, we probably would have had some more innovation. I'm not sure I'm convinced it would be a lot more innovation. And I don't necessarily think it would have showed up in the data per se. Because a lot of this stuff is hard to measure. Right?

Lee: Google is much better than previous search engines, but I'm not sure that actually would have showed up in GDP. It just would have showed up in the fact that users were kind of happier with their online experience. So I'm not sure what I think about the larger economy. I think some of that probably happens. Because it's certainly true that antitrust enforcement has been laxer over the last 20 or 30 years than it was in the mid-20th century. But it's hard to know if the counterfactual isn't... if that's actually an important factor in the relatively low rate of new formation and productivity growth.

Beckworth: Yeah. So, one of the concerns... for example, with Facebook... Mark Zuckerberg recently went before Congress. He was grilled. He agreed to improve their act, to do better job patrolling fake news and other issues surrounding Facebook. This is kind of a standard pattern... is this big player agrees to new regulations that end up being ones that it can probably only serve, or meet, in a cost-effective manner because it is so big, and it ends up keeping out smaller firms. But what you're saying is even if that's true, you're not sure that it'd make that much of a dent in the numbers.

Lee: Yeah, I think it depends on which industries you're talking about and what regulations. I mean, so far I do not think there've been any really big ones. And probably the most significant new regulation of tech recently has been the General Data Protection Regulation in Europe, which was not something that big tech companies per se were pushing hard for. But so far I think that the regulatory environment in Internet technology is pretty favorable. And also in self-driving cars. I've actually been pretty surprised. I was expecting that it'd be a big fight where there were a lot of people trying to prevent those from going on the road until there'd been a lot more testing. And that pretty much hasn't happened.

Lee: So I believe you're calling out Alex Tabarrok on... He had a paper, a few years ago, where he tried to empirically look at can we find a relationship between slow growth and regulation, and was kind of expecting to find a relationship, and found it was much weaker than he had expected. I'm very open to theories about regulation hampering productivity growth. I'm personally pretty concerned about things like occupational licensing as a barrier, but in terms of the general regulatory climate, I have not seen any compelling evidence that that's a significant factor in the overall rate of growth for the economy.

Beckworth: Tim, let's go back to an issue you were talking about a few minutes ago, and that is the effect of this technology upon the labor force. So, let's just assume, for the sake of argument, that productivity does take off. There is massive technological change. And there are people losing their jobs left and right. What type of policies would you suggest? Would you be open to universal basic income? Better retooling of workers? What would be your prescribed policy, moving forward, in a world where there was rapid, more aggressive technological change than we see now?

Lee:  I think it would depend a little bit on what the shape of it is. So, one question, I think, is to what extent it's regional versus kind of formally distributed. I mean, the pattern we've seen recently is that we see amazing economy in San Francisco, New York, the other big cities, and then you have certain areas where there are factories being shut down and so forth.

Lee: And one policy idea that I'm not completely sold on, but I think is really interesting, is an aspect of antitrust, as I understand it. Antitrust enforcement used to be much more concerned with regionalism. For example, they would sometimes block mergers because you had a company headquartered in St. Louis, and it would be acquired by a company in New York, and they would block that because they didn't want all the economic power in... There's often some kind of positive externalities to having... There's services you can provide to a company, and there's good job incentives that are... So one thing I think is interesting is just thinking about kind of place-based policies, and I'm not sure, but I don't think I would go all the way to blocking mergers on a large scale to kind of try to balance that.

Lee: But thinking about that kind of thing, my former Vox colleague Matt Yglesias had an idea I really liked about moving some federal agencies to places. You can move the NEH to Cleveland, say. Because there are a number of federal agencies that really don't need to be in Washington. There's plenty of jobs here. They're driving up property values, but they would actually be very welcome in some cities. So that's one category of things. I'm honestly not sure there's that much.

Lee: And I'm definitely, at a conceptual level, in favor of something like a universal basic income or some kind of income support. I actually like wage insurance, the idea that, you lose your job and take a lower-wage one, that there's a program that will give you some fraction of that difference back. But I don't have a strong opinion about which specific unemployment, welfare-type programs are best. And my hope is that if you have good demand environment, and you do thoughtful policies about making sure that particular areas don't lose their jobs all at once, that you shouldn't need a ton of that kind of thing because most people who want jobs should be able to find them.

Beckworth: So, you're not terribly worried about a future where capital's getting more and more of the income share. There'll be some kind of process through which the distributional issues will be less consequential.

Lee: I think so. I guess I would say that in a world where you get to the really dystopian world, where income equality continues increasing, then I would very definitely be open to more income redistribution. I've become more sympathetic to the idea that the decline of the labor movement has had some negative effects because that's certainly seems to have been a force. But I really do think that the place where you had the best growth of low- and middle-income wages in the last generation or so was in the 1990s, when you had the economy running hot, and I really do suspect that if we had similar periods like that, that you would see some of those tends reverse.

Beckworth: All right. So, in the time we have left, one other policy area... And we touched on this already. You mentioned you would like to see the Fed do a better job managing demand. But what specific policy or monetary regime would you like the Fed to follow? Because you've written on this. In addition to your technology work, you've also written on the right approach for monetary policy it takes. Tell us about that.

Lee: So, I'm definitely on the nominal GDP targeting bandwagon, and I definitely think level targeting is a important piece of that. So I think, in my ideal world, the kind of Scott Sumner program I would definitely sign up for. I'm also somewhat sympathetic to the idea of a higher inflation target. Because we're likely to have a repeat of the 2008 zero percent interest rate situation, and so having a higher inflation target would be another way to do... I think probably NGDP targeting would be a better idea. But, of course, you could also have a NGDP target that's a little higher, which is kind of equivalent to a higher inflation target.

Beckworth: Okay. So, I knew you would give that answer. It's why I asked. Another chance to plug nominal GDP targeting. But I also bring this up because I think that nominal GDP targeting is very useful when it comes to these productivity innovations, to technological change. When it comes to the positive supply shocks, they create movements in inflation and output. They go in opposite direction. So a positive supply shock... let's say more robots, more 3D printing, more smart cars... would lead to potentially faster economic growth and lower inflation. What would the central bank do in that situation?

Beckworth: And I think nominal GDP targeting kind of says, "Step back. Don't worry about it. Just keep demand stable. Don't worry what happens to inflation in the short to medium run. Just keep demand stable." So you don't have to play God. It's just much easier from a policy perspective, from a knowledge perspective, if you were to have these advances in technology. Do you ever think about it from that perspective? Do you ever think nominal GDP targeting is useful from that angle?

Lee: Yeah. That makes perfect sense to me. One other related thing that I think is actually pretty equivalent in a practical perspective.

Beckworth: Okay.

Lee: But I kind of like the idea of talking about targeting average wages as a similar way of getting the same results.

Beckworth: Absolutely. Yeah.

Lee: If your concern is that average people's incomes are going to fall as a result of not being able to find work, if you're targeting people's wages, I think that both may get you to the kind of same policy result and also may be much more intuitive to people. Well, especially with an inflation target. While I think, theoretically, a 4% inflation target has some nice properties, if you tell people, "We're going to make your cost of living go up faster," that's not going to be appealing. Whereas if you tell people, "Our goal is to make sure everybody's income goes up by 2% a year," I think that's something that would make intuitive sense to people and might just make the transition to a system like that a little bit easier, politically.

Beckworth: All right. Well, on that happy note, our time is up. Our guest today has been Tim Lee. Tim, thank you for coming on the show.

Lee: Thank you so much. I enjoyed it.

Support Mercatus

Your support allows us to continue bridging the gap between academic ideas and real-world policy solutions.Donate