Invest in Invention
University of Chicago economist Charles Wheelan and author of Naked Economics: Undressing the Dismal Science and other books talked to Andrew Benedict-Nelson of Insight Labs about what economics can teach us about the impact of coming technological change.
Andrew Benedict-Nelson: In our recent Law 2023 session, we were investigating the consequences of technological advance on the future of the legal profession. We converged on an interesting idea — the real impact of new technologies won’t be on the tasks that any individual lawyer can accomplish, but on the price points of various things the legal industry does. The changes to policy and institutions that result from that would be really fascinating.
Now, when I heard this, I thought, “Wow, this is pretty important for predicting the future not just of law, but of almost everything. Maybe when we try to predict the future, we should be talking to economists instead of tech guys.” So this is what led us to you — I’d like to test this hypothesis.
Charles Wheelan: Okay.
Andrew Benedict-Nelson: So let’s start with the basics. I remember taking the equivalent of an Econ 101 course — supply and demand, rational actors, monetary policy, all that. And I don’t remember technological progress being among those basic shaping factors of the economy that we discussed. Is that the case? Or did I just fall asleep and forget something?
Charles Wheelan: There are a few places where technology enters the discussion in a fairly simple, technocratic kind of way. If you get a technological advance in production, it will shift the supply curve, because it’s cheaper to produce everything. You end up selling more goods at a lower price. But that tends to over-simplify the way things actually happen.
Another area in which you would often see it is in development economics. People are interested in the ways in which new technologies do or do not lead to higher standards of living.
But these are fairly simple, hypothetical curves that you might use in an Econ 1 class. They would have little to no value to understanding how technology actually interfaces with our lives.
Andrew Benedict-Nelson: Okay, let’s go up a notch, then. Let’s say I’m a graduate student who has just been admitted to a PhD program in Economics. Let’s say I’m about to propose some new research in the field. What are some of the key concepts in economics and technology that anyone in the field would assume that I know?
Charles Wheelan: You would probably have to know some ideas from industrial organization. Those are ideas about how firms compete and innovate, as well as the incentives for technological innovation.
You would probably understand some of the policy arguments around intellectual property protection. Economists are rightfully concerned right now about a tension in intellectual property. On the one hand you want to protect intellectual property. If I spend a lot of money to develop some new way of doing things and you can copy it at virtually no cost, the impetus for innovation goes away.
On the other hand, if I can come up with a bogus innovation, that might also turn off the tap of innovation. You see this sometimes with the pharmaceutical industry, where you can pretend to change a drug and extend the patent protection when you haven’t fundamentally changed the formula. Eventually, you want other innovators to replicate technology in cheaper, better ways. So there is a tension over how much you protect innovators.
You would probably want to study something about the promulgation of technology. I work at the intersection of economics and public policy, and we’re very interested in things like the spread of knowledge and best practices, particularly in health care. This happens much more slowly than you would think. In standard economic models, the transmission of best practices is considered to be almost immediate. But in fact it takes a shockingly long time for new ideas to be translated to all of the medical centers and hospitals and so on. So the promulgation of technologies may be a field in which economics needs to look to other disciplines like organizational behavior or sociology.
If you study productivity – the macroeconomic question of how much workers are producing – you would also be interested in the impact of technology everywhere from the shop floor to service industries and even places like education and healthcare.
Andrew Benedict-Nelson: And from an economic point of view, that number also includes automation, right? I feel like that’s a story you hear a lot in our present economic situation — corporations recover but they find out they can replace many of their workers with some form of automation. That’s a story that only makes sense if you take technology into account.
Charles Wheelan: Yes. We haven’t yet talked about labor economics. For that discipline, the substitution of various forms of technology for labor is one of the most important stories of the past 15 or 20 years. Two things happen in that scenario. First of all, the premium for highly skilled labor continues to rise. If you are highly talented, the market for your skills is bigger than ever before — globalization is good for you. New technology also tends to make those highly skilled people even more productive.
But new technology also tends to make low-skilled people redundant. You can see that in any number of changes in our lifetimes — many people don’t have secretaries any more, the ATM has replaced the bank teller, the automated ticket machine has replaced the ticket-taker, etc. When we talk about the labor market politically, we tend to obsess over issues like immigration and trade. But technology by far is the biggest culprit when it comes to destroying low-skill jobs.
So there is no doubt that technology and the labor market interface to produce this environment that you discussed, where firms that are tech-savvy can make a lot of money and the market for low-skilled labor is not great.
Andrew Benedict-Nelson: Here’s why I brought it up — this seems like a common economic problem where technological advance is a key part of the equation. You can’t really think of many ways of automating production that wouldn’t count as a technological advance. By contrast, if we’re talking about supply and demand, I can find hundreds of ways to decrease costs that may have nothing to do with technology.
Charles Wheelan: Right. Let’s talk about another implication of automation. For a long time, the returns to capital and returns to labor in companies were relatively stable. If you’re making cars in Detroit, some amount of the profits got paid out to labor — the workers who show up every day — and some amount got paid out to capital — the shareholders who own the car company.
But when automation replaces labor and profits remain the same, the payout to capital increases. People like Joseph Stiglitz and Paul Krugman have argued that that is a major trend right now, though it’s harder to say how important it will be in the long term. In theory we should become richer as a society, we should create new kinds of jobs, and people should get new kinds of education to fill them. But certainly in the short- and medium-term, the story of technology is that it is good for owners of capital, good for highly-skilled people, and not particularly good for low-skilled workers.
Andrew Benedict-Nelson: I’ve got another basic question about economics and technology. There are plenty of frustrating little mysteries in our everyday lives that economists are good at explaining, like the price of gas. Can you think of any common problems in the development of technology that are a lot easier to understand if you have an economic perspective?
Charles Wheelan: Yes, but economists aren’t terribly good at conveying their answers. …
Let’s look at something like carbon emissions, which is an outcome of new technology. Economists would explain this as a classic case of an externality. In about week four of your microeconomics class, you would learn that markets work pretty well except in cases where you don’t bear the full cost of your behavior. Pollution is a classic case where you and I could both be producing something, but neither of us has to pay for the cost of the pollution. Therefore, we’re both going to pollute too much. If we agree to stop polluting, it may benefit one of us more than another. If we don’t all agree, it may give one of us the ability to steal the polluting industries away. So economists can explain pretty easily why it’s so hard to solve this problem that emerged from new technology.
Health care provides another example. In health care, many of the increases in cost are driven by new technology. In week five of that same microeconomics class, you would learn about moral hazard — when you stop bearing the cost of your actions, you stop caring about the price of your behavior. Because most people have someone else paying for their health care, they aren’t sensitive to price and naturally demand the best and the newest and the fastest.
So think about technology in that equation. One thing that technology can do is make things bigger and better and faster. But another thing it can do is make things cheaper, cheaper, cheaper. It can give us DVD players that cost $8. In a field like health care, though, there is virtually no demand for technologies that decrease cost. When someone else is paying the bill, we naturally demand things with more bells and whistles. There is no one to do what Toyota did in the automotive industry, which is to innovate by taking all the existing things that cars can do but just making them better and cheaper. A Corolla wasn’t a better car than a Mercedes; Toyota just brought the price point down.
But there’s no one who is incentivized to create “Corolla” health care. There’s just a bunch of people who keep promising to buy you a Mercedes. There are a few notable exceptions such as the Veterans Administration — because they know injured veterans will be in the system for their whole lives, and they also have a fixed amount of funding from Congress, you see some innovation on the cost side of the equation.
Andrew Benedict-Nelson: There’s something really important in what you said, and I want to pull it out and emphasize it. You seem to be saying that from the point of view of economists, people are only going to invent the technologies that they are incentivized to invent.
Now let’s stop and think about that. We tend to think that technological progress is this process that is humming along all the time, just churning out new machines. But what you’re arguing is that those technologies take forms that are pretty specific to the sort of market around them.
Charles Wheelan: That’s the nature of entrepreneurialism. You anticipate a need and then attempt to meet that need.
Andrew Benedict-Nelson: Does this also suggest that in the long run we can use economic levers to control the inventions we’re going to get? That sounds sort of paradoxical.
Charles Wheelan: It’s not paradoxical. That’s why the Soviet Union could launch Sputnik before us. They had a centralized system, so if Moscow decided that they wanted to throw a tremendous amount of resources at a particular problem, they could do that. There’s no market demand for putting a dog in space. The downside is that they couldn’t produce tennis shoes or mouthwash that anyone would actually want to buy. But if we decide that we want to go to Mars in 15 or 20 years, there’s no question that we could do it.
Andrew Benedict-Nelson: Right — from an economic point of view, it’s just a question of what “we decide” means. Surely with all the trillions of dollars in the world economy, you could get the resources necessary to do it. But someone actually has to aggregate them in the right way.
Charles Wheelan: Right. So is it central planning where we forcibly tax people to fund the enterprise? Will we decide in some market-based way? Probably not, because you could observe there is only a relatively small private market for space travel compared to the likely cost of the research. So you can reasonably predict that the private sector is not where the innovation needed to get to Mars will occur.
Andrew Benedict-Nelson: Just to round this out, there is also an economic story that explains the value of government investment in basic science, isn’t there?
Charles Wheelan: Basic science has the attributes of what’s called a “public good.” Let’s take a basic example like an expedition to go around the world and prove that it’s round. That discovery has pretty limited commercial applications, because while the initial costs are enormous, the costs of using the information once it has been discovered are basically zero. You can’t go around telling people, “Hey, I’m only going to tell you whether the world is round or not if you pay me $100,000.” Word gets out. The same is true for other basic discoveries like the fact that germs cause disease.
So there is very little incentive for private firms to engage in this kind of research, even if the potential results are really powerful. It will usually be under-provided in the absence of some kind of public support.
Andrew Benedict-Nelson: Okay, so I feel like I now understand some of the basic concepts economists use to think about technology. Let’s get to the problem at hand.
In our session, we became really interested in scenarios where the effects of new technologies grow out of their impact on price points. Let’s say you’re about to give a lecture on one such scenario from an economic point of view. What are the economic concepts I need to have in my mind to make sense of this?
Charles Wheelan: You would want a basic understanding of marginal cost curves, which is some of the stuff we talked about earlier. So if all of your inputs are getting cheaper and demand stays about the same, you can assume that you’re going to sell a lot more of these goods at a lower price.
But beyond that, the analysis isn’t terribly sophisticated, because there are all sorts of interactions between price points and the cost curve that you can’t easily predict. So the kind of analysis I just described probably explains PCs pretty well. Every time I go to buy a new computer, it costs roughly the same amount. But the amount of computing power that I’m getting for that price gets better and better. Makes sense. We could predict that. But what we could never predict is that Apple would use the same decline in price points to produce new kinds of products we had never imagined and make a gazillion dollars.
So you can be Dell, following this relatively predictable path where things get cheaper and cheaper and smaller and smaller, or you can be Apple and create the iPod or the tablet. I think you would certainly rather be Apple. But economics doesn’t capture how to do that very well at all.
Andrew Benedict-Nelson: Earlier on, you basically drew an equal sign between “basic research” and “public good.” Is there an economic concept that maps just as well on to the decline in prices driven by new technology?
Charles Wheelan: Not one that I can think of. Most of what’s out there tends to be so formulaic as to be useless. There is this interaction between delivering new features, more quality, etc., with the fact that you can produce whatever you’re producing cheaper than ever before. There’s no easy way to predict which path consumers will prefer more. Do they want the same laptop for a lower cost or a faster, more powerful laptop for the same cost?
We can see that there are some industries, like health care, where no one ever wants a cheaper version of the old laptop. That’s a function of incentives. I would think that most of the time people would want some combination of the two — but again, it’s hard to say what that means.
Andrew Benedict-Nelson: Yes, I’m beginning to see how hard this is. Think about netbooks — these fairly inexpensive laptops that you can only really use to navigate the Internet. If you look at the technical specifications alone, they are a predictable point on the curve. They are equivalent in power to older laptops, but much cheaper. But the thing is, there wasn’t a market for machines like that until we also had this completely independent phenomenon, which is widespread Internet access. So the shift in price points only created an opportunity; it wasn’t the change itself.
Charles Wheelan: Right – on its own, the cost curve doesn’t predict that you could move your hard drive to the cloud. That’s just the sort of thing I’m talking about, and it’s really difficult to model.
Andrew Benedict-Nelson: Okay, so we don’t have a standard model. But I’m guessing there’s still a lot that we can learn from economics. Let me frame this problem in a different way.
One of the earliest novels to appear in English was Robinson Crusoe. Since then, we’ve had thousands of other stories about people getting stranded on islands. If I give you the scenario of a person stranded on an island, you’re going to have a lot of trouble coming up with a plot point that hasn’t occurred in at least one of these stories. We have a general idea of what we can expect from the genre.
Let’s look at the scenario of technology leading to declining price points in the same way. We concede that there is no one model that explains everything. But what are some of the phenomena that we might expect to encounter in this scenario? Are there any examples from your own work that could guide us?
Charles Wheelan: Well, I’m a policy guy at heart. I’ve got one from the world of education that isn’t exactly about price point, but it is about technology. In the education policy realm, the technology that has the potential to disrupt K-12 education reform is called “value-added testing.” Test scores are generally a very poor metric of teacher quality. You never really know the inputs — some kids arrive ready to learn, some kids don’t. And you can’t even attribute all of the gains that students make to the teacher.
But there has recently been a lot of innovation in “value-added testing” that allows us to control for things like past performance and socioeconomic status. So we have a better gauge of what the teacher actually adds. The change that comes out of that would be completely new mechanisms for hiring, firing, and paying teachers. We’ve been stuck for years with this model where you’re paid based on how many years of experience you have or whether you have a Master’s degree — neither of which seem to correlate with performance very well at all.
So the conclusion is that technology has opened up a new and different realm for how we can get better teachers in the classroom.
Andrew Benedict-Nelson: It seems to me that in your scenario, there actually is an impact on price. Because you would have much better information, you would create a new market along the lines you described earlier — better teachers would become more expensive, mediocre teachers would become cheaper or even redundant. The point is that the role of technology would be to end asymmetrical information.
We talked about this in the Law 2023 session too. It’s kind of like buying a used car. Today, I can go in to a used car lot armed with quotes on the exact same make, model, and year from dealers all around the country. Access to that information prevents me from overpaying.
Charles Wheelan: eBay operated on a similar principle. Sometimes the real technological innovation is just making markets bigger. You could have sold many of the items on eBay before through the classifieds, but if you didn’t think you had a good chance of finding a local buyer, it might not have made sense to pay for the ad. So yes, information can drive the change.
Let me give you an example that actually works the other way, though. In medical technology, we have more kinds of imaging and tests than we’ve ever had before. But these tests are not perfect, so when you’re testing healthy people, the chance for false positives is high. Historically, we didn’t test people until they were symptomatic, so the chances for false positives was low.
So let’s say you get a scan for potentially cancerous tumors. We could find an early-stage tumor that, 50 years ago, no one would have found until you were dying of it. We could save your life. But there’s also a good chance that that tumor is completely benign, the equivalent of a spot on your liver.
The problem is that once you’ve seen the spot, you are going to do follow-up procedures. And some of those follow-up procedures are quite dangerous. Just going to a hospital is potentially dangerous, given the number of hospital-acquired infections. And you even have public health officials pushing back on some of these screenings, saying that it might not be a good idea unless there is a hereditary factor or some kind of symptom.
So this is a situation where the information provided by technology actually has ambiguous effects on healthy people.
Andrew Benedict-Nelson: So I’ve got an idea. I like this notion that the real value that economics provides us is a sense of what to look out for in this crazy, unpredictable mix of new technologies. To go back to the Robinson Crusoe metaphor, it may be that economics is the thing that helps us level up from saying, “Oh, there’s some fresh fruit and there are some wild pigs and there are some fish in the river” to “Oh, here are some potential food sources, let’s compare them.” You need a more general concept of value in order to recognize that.
If you play it out, that’s really important for recognizing opportunity. So let’s say I have a technology that makes any given person’s commute one percent cheaper. I think an economist could easily prove to me that most individuals will not notice or care. But they might also be able to show me who does have incentives to make all commutes one percent cheaper, like the local transit authority.
Charles Wheelan: Right. Once something has happened, economics is pretty good at thinking through the ramifications. It can tell you how it might affect competing products or how it could reduce the cost of a manufacturing process. You could probably model what would happen to the market for people whose skills are required to make the technology. We can think about almost anything that involves a general equilibrium model. Even if the actual models stink, the process of modeling is valuable.
Andrew Benedict-Nelson: I feel like there are actually a few ideas from the tech sector that match up with this finding really well. They go under the names of “agile development” or “lean startup.” So let’s go back to the netbook example. At some point, the main question of whether netbooks would be a viable technology or not depended on the market reaction and nothing else. The “lean startup” idea suggests that instead of doing a lot of consumer research and surveys and such, you should just release the product and improve it over time based on market feedback.
What our Robinson Crusoe metaphor suggests is that economics could help us do this is a broader way. It could provide a diverse toolkit that helps us assess what happens when various technologies hit the market. And you could presumably monitor for things besides “Are people buying our product?”
The point is that you wouldn’t have really finished “inventing” your technology until you introduce it to the market.
Charles Wheelan: And you may have to introduce it in a big way. Look at a company like FedEx. They couldn’t have figured out if there was a viable market for overnight shipping if they only introduced it in two towns. The whole point of the service was that you could get it anywhere.
Andrew Benedict-Nelson: Hmm, I’m sensing that this may be a problem when it comes to innovation in the legal field too. Let’s go to an extreme. Let’s say we had a machine that could give everyone access to a lawyer for a dollar. Well, we couldn’t meaningfully test it in just one town — the nature of the Law is that it’s bigger than just one town.
Charles Wheelan: This also raises the issue of regulation and innovation. So in India, you don’t have much of a textiles industry, because the law makes it very hard to fire anyone once you have more than 99 employees. As a result, most of the firms tend to be 99 people or less. But innovation in textiles requires scale. This helps explain why so much of the textiles industry has gone to China and not Southeast Asia.
So let’s look at innovation in the legal industry in the U.S. in the same way. You may have similar limits on innovation that result from the fact that it’s hard to combine law and consulting or law and private equity.
You know, if we were completely unencumbered by regulation, my first thought is that I would like to see a lot more innovation in services. I would love to be able to go one place that could provide a business with capital and legal advice, for example. But there are regulations that preclude that kind of innovation.
Andrew Benedict-Nelson: That’s important for the conversation about business models too. For example, I did my taxes on TurboTax this year. When I got to the end of the process, TurboTax asked me if I wanted to sign up for Mint.com, which is run by the same company. Well, one of the reasons that business move was possible for them was that technology allowed them to know the exact moment at which I completed my taxes and might be thinking about how to allocate a refund.
So you could imagine that there are potential businesses and technologies in the legal profession that are being held up because regulation does not allow them to line up services in that way. That’s not to say those aren’t good regulations — I have no idea.
Charles Wheelan: Well, often they’re not. I did my dissertation at the University of Chicago on rent-seeking in the regulatory process. A lot of times this is just professions or other groups carving out fiefdoms to prevent competition. …
Andrew Benedict-Nelson: Do you have anything else you’d like to add from an economic perspective about the future of the legal profession?
Charles Wheelan: I think it has to be something with the bundling of services. For example, I have a lawyer because I have a literary agent. It is extremely helpful to me that all legal services that I get — reading contracts, writing contracts — come bundled with that literary agent. I don’t pay by the hour — it’s just part of the percentage that I pay. It’s highly specialized — I know that contracts for authors is all that they do. And I don’t have to interface with the lawyer directly.
So if I were a private equity firm, I think that having legal services embedded within my investment bank would be extremely helpful. I think the other thing that you will see, just as we have seen in health care, is more resistance to fee-for-service and more innovation in billing. I think you are going to have more situations where you pay for something like a real estate closing and building in incentives for it to be cheaper, faster, and better.
Andrew Benedict-Nelson: There’s an interesting corollary to the idea of bundled services. Imagine a world in which we primarily think of lawyers as components within other services that you buy. They would be like batteries. Sure, we’ve all bought batteries. We may even have a favorite brand of AAA batteries… Energizer, Duracell. But most of the batteries that we use are already built into to various devices like our computers and our cars and our phones. And we judge the total performance of the device. So if I have a phone that is dying all the time, I say, “This phone model has bad battery life” rather than “Oh, I wish my phone-maker had chosen a different brand of battery.”
The point is that if lawyers really were integrated into all of these services, we might arrive at a very different ideal of what the legal profession is. It might be that most people never actually have any sort of economic exchange with a lawyer, for instance, just as I buy relatively few batteries in life.
Charles Wheelan: I know it drives me crazy when batteries are not included. I know I’m willing to pay a premium to make sure there will be batteries when that toy is opened on Christmas Day.
Andrew Benedict-Nelson: So “lawyer included” could be not just a convenience, but a new source of value.
Charles Wheelan: Right. In the end, it’s about making life easier for people. That’s what Apple did. Apple products certainly aren’t cheaper. But there’s a feeling that you just take it out of the box and it works.
Andrew Benedict-Nelson: Right. Apple incorporates all sorts of technologies into a device like the iPhone, and the beauty is that I never know about any of them.
You know, this may actually be the most interesting outcome of our technology discussion. So think about the history of cars. Once upon a time, if you wanted to add some component to your car, you had to find and buy it and install it yourself.
Charles Wheelan: Right — I grew up in the era when you went and bought the car stereo yourself.
Andrew Benedict-Nelson: Well, one of the reasons why it’s become easier for car manufacturers to create more integrated offerings is because of technological progress, including changes in price point. It’s easier to offer something like smart phone integration after you know that lots of people have smart phones, that you can easily find engineers who understand smart phones, that smart phone interface components are cheap, and so on.
You could see the integration of lawyers into other services happening the same way, if the legal industry is innovative enough. You’d have “lawyer inside” just like you have “Intel inside.” They’d be providing all sorts of value even though you would never know they’re there.
Charles Wheelan: It’s not hard to see the appeal. I’m not a small business owner, but my wife was. To have one-stop shopping for all of the things you need to do to open a small business would be an enormous help.
Andrew Benedict-Nelson: It’s so bizarre — what we’re ultimately talking about is treating lawyers themselves as a kind of technological component that follows the same economic rules we’ve been talking about.
Charles Wheelan: Ha, yes. But I think that’s a good way to think about it.