Shane Greenstein on Innovation, the Internet Age, and the Future

Scott Wallsten:

Hi, and welcome back to Two Think Minimum, the Technology Policy Institute’s podcast. Today is Monday, May 3rd. I’m Scott Walston, President and Senior Fellow at TPI. I’m joined by my cohost TPI Senior Fellow Sarah Oh, and today we’re thrilled to have as our guest, Dr. Shane Greenstein, Professor at Harvard Business School. 

Shane is the Martin Marshall Professor of Business Administration and Co-Chair of the HBS Digital Initiative. He also co-directs the Program on the Economics of Digitization at the National Bureau of Economic Research, and because HBS provides such modest bios, I thought I would embellish it a little bit with just some numbers from his CV. 

So, Shane has written nine books and edited volumes. 71 articles in peer review journals, 47 chapters in books, 15 published proceedings and reports, 52 invited publications, 150 articles for IEEE, which means that we could probably calculate the opportunity cost of your being with us today in terms of lost articles.

But we’re thrilled that you chose to bear that cost, and we hope we make it up to the rest of the world for having you. So, thanks so much, Shane. 

Dr. Shane Greenstein:

Thank you for having me. It’s no cost at all. 

Scott Wallsten:

Well, then what does that say about what you’re normally doing? *Laughs*

So, let’s start with one of your more recent articles, which is titled “Interpreting the Origins of the Internet,” which really follows from a lot of your research on the history of the internet and computers and so on. And you say, basically, when you tell us about it, you say that a lot of the story… a lot of conventional wisdom is correct, but a lot of it isn’t, and the part that isn’t are the parts that explain the Internet’s importance.

Dr. Shane Greenstein:

Well, there are a couple of different myths that you see commonly. I have to say part of the motivation for that article was [point] A, nobody read my book. Maybe you read it, but very few people wanted to read 500 pages, but you might get them to read less. And the second thing that motivated this, a typical audience I speak to now is under the age of 25, and they were born after the internet deployed, and they inherit these common myths. 

So, one of the common myths is DARPA invented it and then it diffused. This is a sort of a schoolboy version of invention. You know, someone imagines the new product or service, and then it’s thrown over the fence into commercial markets, and firms sell it. That does not describe the internet at all. It’s not a single invention. It’s the cumulation of many inventions over a wide period of time and coming from a set of both users and producers, so that the inventions themselves have a mix of motives, everything from commercial to non-commercial motives, and they accumulate over time. So, they’re not anticipated by… the earliest inventors don’t anticipate the accumulation. 

It would be completely inappropriate to say there was an individual at the very beginning who recognized what the architecture would look like 20 years later. There wasn’t such an individual who designed the whole architecture. Despite the computer science community using the word architecture, they don’t mean it in the same way that we mean when we look at a building. So, that’s one myth.

Scott Wallsten:

Actually, before you go on. So, they use the term architecture differently than most people use it. So, what’s the difference?

Dr. Shane Greenstein:

You talk about processes for interplay. So, if you want to think about an API, if you know what an API is, a boundary between a software, they think about procedures and processes between different software protocols or different modules of software. So, they think about how a system operates when you put together a lot of different parts, and so architecture for a computer scientist is thinking about all the various different processes and the interplay between them. I’m glad somebody thinks about them. Don’t get me wrong. That’s a really important thing to think about, and certainly one of the important innovations in the internet was to put the intelligence at the edges rather than in the center. That was a central idea, quite different from the telephone network, which puts the intelligence at the switch and doesn’t put much intelligence at the edges. So, by putting it at the edges that allows our computers and our devices, and what we now recognize as your laptop, your smartphones, everything else to include a lot of computing and then do something with it and translate into a form that could be communicated with another device somewhere else. That’s the architecture in the computer science sense is a set of rules that permitted that to occur. That was quite clever. That’s a wonderful thing. So, that’s one myth. 

Sort of another myth kind of comes in the other direction. There’s a lot of versions of this. There’s the Silicon Valley myth, which is that Silicon Valley invented the internet. But there’s another version of it, which is just that commercial markets invented it and invented everything, and that’s also untrue. It could not have been done without public assistance and it could not have been done solely inside of bell outfit. You know, a few firms in Palo Alto. 

The history tells you that the ideas came from a mix of mainstream firms and iconoclasts from all over the place, and the really striking thing is how many contributions come from all over the United States and really all over the world today, but at that point, all over the United States. Then the other striking thing, which I was just really blown away by when you get into details is the public private mix. So, the things like Wi-Fi, you know, something we take for granted today, the initial standards come out of the IEEE, which is a voluntary organization, but the first commercial product is an Apple product. You know, Michael Dell’s rivalry with Steve Jobs has placed us in the important role and getting a second Wi-Fi product in the PCs. And that’s that mix of public and private, and just that example that you find all over the internet. 

Another one is Apache, initially comes out of the universities. It’s shareware, there’s no price on it. There’s no license on it. Then it adopts the open source licensing regime. It becomes the most commonly used web server in the world. It doesn’t show in GDP directly because nobody sells it. There are millions of copies of it everywhere. It supports the internet. That’s coming from a very different place than the commercial market.

Scott Wallsten:

Tell us, before you go down too far, a little bit about what was the public aspect of it, because you started off by saying that part of the myth is that DARPA invented it and that that’s not true, but then you say that…

Dr. Shane Greenstein:

Yeah, DARPA played a pretty important role in being a catalyst.

Scott Wallsten:

Were they the main public component of it? 

Dr. Shane Greenstein:

Yes, initially. Without question initially, yes. Just as I think it’s also accurate to say DARPA, just going to the present, DARPA’s catalytic role in autonomous vehicles, it has to be recognized, right? The DARPA Challenges that organized a number of R&D efforts into actual… You know, and then they put together a course and that focused a lot of engineering effort, and so give DARPA it’s due. It’s pretty important that it played this catalytic role. You know, not everything DARPA did worked, but it’s also a schoolboy myth to think that, you know, that eventually alone is what got us the internet. You have to get deployment and use, and learning from use, and you had to get national deployment, large number of other things. 

The other thing to also say a little more subtle for Washington, but this is a Washington-based…let’s say it, the crossover, the handover from public funding to private funding traditionally has been an awkward moment, and that’s because private firms can see their interests. Once they can see them, they’ll try to influence the public decision-making, and in many countries, that’s a very difficult handoff to make. 

And there was a myth that the internet was clean and everything worked well, and that’s just a myth also. It was a very difficult handoff from public to private, and we were fortunate, I think the right way to say this is we were fortunate that honest people got involved and kept it an honest and uncorrupt process. 

Scott Wallsten:

What were some of the tensions, and what risked being corrupted?

Dr. Shane Greenstein:

It’s possible to commercialize the internet and preserve a monopoly behind it. So, we could have been monopolized to network that would have been quite distortionary on later events.

Scott Wallsten:

Where things like AOL’s…

Dr. Shane Greenstein:

IBM or AT&T. AT&T fortunately wasn’t interested. IBM was. To give an example. 

Scott Wallsten:

Was AOL’s closed garden a risk or an innovation at the time?

Dr. Shane Greenstein:

At the time, it was regarded by many technical users as a risk. I would think in retrospect, I would see a lot of benefits in AOL, and that it did what lot commercial markets do in that it translated what were obtuse and maybe difficult technical processes into something that was a little easier to handle if you weren’t technically skilled, and AOL played a pretty important role in making parts of the internet accessible for the general public. So, I think in retrospect, I’d say it’s pretty positive.

Scott Wallsten:

What about looking forward? I mean, what are the implications for how we think about R&D policy? And also you talked about tech transfer that tech transfer is more than dumping it over on the other side of the fence, and we’ve had lots of policies and rules related to tech transfer. It seems like something we’re always thinking about and always getting kind of wrong and kind of right. So, what does the development of the internet tell us about how we should work on these two things going forward, and in particular, it seems worthwhile to think about now because with all of the coming spending, a lot of it, presumably is going to be on types of R&D.

Dr. Shane Greenstein:

Yeah. So, an example I have become increasingly fond of is autonomous vehicles. Autonomous vehicles today is a lot like the internet of 1985. It’s inevitable. It might come soon. It might not. It might take a commercial form that is monopolized. It might take a competitive form. A lot of things being done at the federal and local level could make a difference to the way the autonomous vehicle commercial market plays out. So, do we see lessons from the internet that play a role in that? 

So, I’ll give an example. If you talk to someone in the industry today, they’ll tell you immediately, autonomous vehicle on-tracking is going to deploy in the next couple of years. At least in some sort of prototype, and arguably at scale. Something that would help them a great deal are paints on the road that the LIDAR can see, not just white paint. It’s pretty straightforward to put some chemical into the paint that the autonomous vehicles can detect. That has major consequences for safety, and that has a public benefit, and it isn’t going to happen unless there’s a public decision to make it happen. That’s a simple example, but there are lots of other examples having to do with autonomous vehicles that are very analogous to the internet. Public standards would make some things easier, particularly when it comes to vehicle-to-vehicle communication, and that’s not going to deploy for 10 years, but we need that. We were already getting autonomous vehicle use in agriculture and in mining where it’s possible to do things in geo-fenced areas, and again there, there’s an enormous public investment. The GPS is being used to get the accurate… You don’t want a tractor driving off the road. You don’t want a mining vehicle, again, driving off, you know. It would surprise nobody that spectrum plays a role in this and the spectrum that’s available. 

Scott Wallsten:

So, let’s go back a second to the paint, because that seems like a useful model for thinking about it because presumably, I mean, you’ve got lots of different places that would need to agree. There’s some kind of standard involved, I would assume, and it seems pretty simple, but you seem to be suggesting that it’s not going to happen anytime soon.

Dr. Shane Greenstein:

Well, it’s not simple because there’s private firms. You know there’s… let me say it this way. Have you seen, have you heard recently about the Department of Defense and their attempt to get a little bit of cloud infrastructure?

Scott Wallsten:

Yeah, I’d heard something about that. They at least had a great name for it. 

Dr. Shane Greenstein:

Yeah, and we’re going to get the same kind of scale issues in paint. It is going to be in the billions of dollars. It is going to be a choice with direct consequence for the financial health of firms. It’s going to have to use public procurement standards and norms. It’s going to be difficult to do in a way that actually gets it all done without it getting tied up in administrative legal hearings.

Scott Wallsten:

And what do you expect? Is it going to happen? And in what time, period?

Dr. Shane Greenstein:

I don’t know. I don’t know. I that’s why when I look at the internet, I say we were fortunate. We had, in retrospect, we got a pretty good process.

Scott Wallsten:

Do you think it was partly because nobody knew what it could be? I mean, with the paint…

Dr. Shane Greenstein:

That’s incorrect. Insiders knew, for sure, and there were attempts by the firms that knew to try to influence it. The general public didn’t know, but.. 

Scott Wallsten:

And so, you attribute its success… Sorry, this is the wrong way to say it, but you said we were lucky that there were lots of honest people involved. So, do you really attribute this to the individuals who were there rather than sort of the economic incentives facing the various firms involved?

Dr. Shane Greenstein:

Yeah. Actually, I had coined a phrase for it, the “honest policy wonk.” It’s a little cheesy, you know, maybe corny, but I’ve met a number of public servants who are really honest policy wonks, and they do… it is pretty important that you have public servants with character. So yeah, actually it did matter. It did matter for the internet. There had to be a change in the National Science Foundation Charter. That was an act of Congress. There was, you know, Congress actually had to act. The National Science Foundation, during this procurement, came under investigation, but those investigations played a really important role because it kept the scrutiny up and the scrutiny generated a change in what NSF did. And NSF, as a consequence of that, put up for a national exchange points and that had positive impact on the interconnection for the backbone of the internet. So yeah, I mean, concretely, it did matter. I mean, I, you know, trace the direct links of positive characteristics that the internet inherited at its earliest moments from the way it was commercialized. 

Then there were a couple of errors. I think we’re still living with them that we recognize today. There was no national measurement set up, and so, you know, we may get into cybersecurity in this conversation. We still, as a country, we still have a problem in that we have no national network measurement framework. We have a lot of, as is typical for the US, a lot of fragmented efforts and fragmentation is good at solving local problems, and so rich firms set up a lot of processes to keep their own networks secure, and that’s great, but we don’t have anything at the national level other than for defense.

And so, we pay this price in that we have a lot of IP addresses that are used for spoofing, and we don’t necessarily have a lot of thought going into the redundancy of the present environment, issues over encryption and insufficient R&D for cybersecurity. These are all a consequence of the way we commercialized it. We kind of just sort of put it out there and didn’t think it through, and different governments have put efforts into this. I mean, it’s not like it’s been that it’s been complete neglect, but we still pay for this. 

Scott Wallsten:

Is there a good model you can think of that doesn’t involve an authoritarian government? There are countries that deal with this, and they may not be the ones we want to model after. 

Dr. Shane Greenstein:

No, it’s not the model we want. Can I say it to you a different way? Because your question, I actually think that your question indicates a common assumption that lack of coordination is a weakness in governance. I’d say it slightly differently. Just as having initiative from lots of different commercial players over the long run is rather good for the internet because it helps us sample from a portfolio of options. In governance and in policy, it’s actually a useful characteristic of the United States that we don’t have a single coordinating policy regime that then accentuates a really bad decision once we make one, right? It’s actually a pretty good characteristic in the long run to sample from multiple sources. Having said that, it’s a mess in the short run. 

Scott Wallsten:

You’ve given us the cost and the benefits of this approach. 

Dr. Shane Greenstein: 

And in the short run, we get these places where we neglect things, and that’s a byproduct of having that kind of governance system in the US where one part of the system works really well while another part doesn’t. Antitrust will be asleep for a while, while the other part is really good on universal service, or universal service is asleep for awhile and antitrust is really vibrant, and so you get this kind of back and forth

Sarah Oh:

From your work on the origins of the internet, do you consider the internet exceptional? Because I would say it’s like a once in a generation type of emergence, the internet. But can we have implications from the origins of the internet on future R&D, or is it like a once in a century kind of confluence of events? Are we going to have another internet type of invention? Can we encourage it, or is it like a big bang moment?

Dr. Shane Greenstein:

I avoid the word exceptional. I prefer the phrase historically unique. So, the uniqueness identifies something you’re after, which is correct. It’s technically unique, and we all lived through a point in time, and it happened in a certain place, at a certain time, and for certain reasons. Because, I mean, we can go into it a little bit, because the US had a set of policies that allowed commercial iconoclasts to experiment, and you know, what made the United States communication market different, particularly from almost every other country, and so, we got frontier experiments here earlier. Having said that, right, that’s historically unique. There are things about it that are not exceptional, that are very general. Having a commercial market that’s open to experiments makes the US… that’s a very general statement. It still applies today. I still think it’s very valuable to have multiple experiments in artificial intelligence, which some people regard as a once in a generation thing or autonomous vehicles is another once in a generation, or drones, which is another once in a generation kind of technology. I think experiments are a good thing, open to multiple points of view, and having markets that have open distribution, or let’s just say a distribution that’s not got heavy barriers being brought by established firms. So, it’s open to competitive alternatives. Those are general features. 

Another thing that I don’t think is particularly exceptional about the internet is eventually revenues have to cover costs, and early on in the internet, there was this point of view that you didn’t have to. That you could have the dot-com boom, and then you could start a business, and somehow the revenue would magically appear at some point in the future.

Scott Wallsten:

Those were great days in terms of transfers from ultra-rich venture capitalists to upper middle-class people.

Dr. Shane Greenstein:

Actually, just society as a whole. There is a theory, I have to admit, I kind of liked this theory because it does fit that event. Although, I’ve never looked more broadly that booms, these kinds of crazy booms weave these capital investments that have long lasting consequences that can be positive. I mean, they can be negative too, if you put it all into bridges that end nowhere. In the case of the internet, that boom left North America with quite a substantial amount of backbone investment, which was positive for the urban areas that were connected to it. I would say that. 

Scott Wallsten:

There are also businesses that couldn’t succeed at the time that turned out to like cosmo.com where you can have that candy bar delivered web van.

Dr. Shane Greenstein:

Yeah, I and web van and so on. I mean the bad part. I mean, the part that is very general is that Wall Street, you know, needed a little bit of reform, and it was pretty apparent. Anybody who was watching could see the information was not being transparently put out there, and we pay for it again, seven, eight years later when we had another financial crash. 

And let me say it a slightly different way. Let me say it in common language. Financial markets are really bad at saying no, and it’s well known that the analysts whose job it is to say, “Don’t invest in this firm,” come under a lot of pressure not to say that. That’s why it’s hard to be an auditor, and it’s hard to be an analyst who says this firm is not meeting its promises and they’re going to lose money, and the firms involved directly try to suppress that information. That’s an old idea. 

And so, you get the Worldcoms of the world in settings where you don’t have good auditing, and you get dot-com bubbles in settings where you don’t have independent information to, for better or worse, bring out the skeptical point of view. So, when you look at the institutions in Wall Street in the late nineties, that was actually a problem. The dot-com boom exposed that, and there wasn’t much done about that afterwards, and so as a country, we again paid for it seven, eight years later, when, again, lack of transparency in now much more complicated financial transactions in the subprime mortgaging market ended up again, catching up with us. So, that’s, I have to say, that’s still an issue that we have to think about.

Scott Wallsten:

Although, you know, again, if you’ve said that both the costs and the benefits of that feature of the economy, because there were benefits that came out of the dot-com boom. Of course, it’s an impossible counterfactual, but what would the world look like without that, if we hadn’t had the dot-com boom?

Dr. Shane Greenstein:

Right. We would have, we’d have really mediocre email.

Scott Wallsten:

Well, I think we do actually.

Dr. Shane Greenstein:

Yeah, yeah. I guess the internet would be nothing but email. No. I mean, what came after the dot-com boom, I mean, historically it was web 2.0 and 3.0 broadband, and that was going to show up eventually. By the way, that’s the other third thing that happens that’s very durable, that’s not exceptional. That things diffuse over time, new technologies don’t get used immediately and never, this just never happens, and then you get historically, you get what we still have today. I often think of it as the optimistic versus pessimistic view. So, let’s characterize 20 years just to say it. So, the optimistic view is new technology shows up. It shows up first in the richest cities with the richest users, with the wealthiest users and the most educated users, for the most adventurous, and the optimistic view then says over time, that becomes easier to use and it becomes more standardized and cheaper, and so it becomes more widely available. So, that’s a description of broadband that starts in San Francisco and New York, and then deploys in many, many, many cities around North America. 

Okay, now there’s the pessimistic view. So, the pessimistic view starts in the same place, by the way. The richest and most adventurous users get access to the new technology, and they always stay ahead that way, and they’re always getting access to the new ones, and suppliers don’t see much of a reason to ever direct their attention elsewhere, and then if I can use the word Marshallian externalitiesright? So, Marshallian externalities then reinforce those benefits. So, the bigger cities get better. Suppliers are better. Firms’ users want to be near them, and workers want to be near them, and so it becomes a self-reinforcing cycle. That’s sort of a pessimistic view, and that says 20 cities in the US, if we’re lucky, get access to the frontier technology, and everybody else has left behind. 

And so, you know, your argument for that one, the standard examples for that are, again, the best backbone service, the best wireless service… for a while there, I thought data centers were going to look like this because of the economies of scale, but data centers are starting to show a lot of indication that they’re going to spread pretty far.

Scott Wallsten:

So, where do you think we fall on the optimistic versus pessimistic view 

Dr. Shane Greenstein:

Today? 

Scott Wallsten:

Yeah, today.

Dr. Shane Greenstein:

In what technology in particular?

Scott Wallsten:

Let’s say, let’s say the things that people rely on most heavily right now. Let’s say education, medicine, and entertainment.

Dr. Shane Greenstein:

I was going to say automobiles. Automobiles is a good example. Level two autonomous vehicles are available everywhere in the US, right? So, that’s because the automobile distribution is pretty wide. So that actually fits the optimistic view. You can live almost anywhere in the US and get a good truck. Eith, you know, for a mainstream vehicle, you can get a cruise control and little beepers that will tell you if somebody is driving close by.

Scott Wallsten:

You can check your blind spot anywhere in the country.

Dr. Shane Greenstein:

Yes. You can check… By the way, that’s the optimistic viewpoint, right? That’s something that was invented 20 years ago shows up in every common…

Scott Wallsten:

And that’s actually not a small thing.

Dr. Shane Greenstein:

Right, it’s a good thing. In terms of the internet, it depends if you regard 90 to 95% as an optimistic or pessimistic view. What we see in the data there, you know, something like 80% of households and certainly a higher percentage of business have access to a broadband network. The data suggests we can get to 90%, if you’re willing to accept wireless as a way of accessing the internet. Then we get into the 5 to 10% of the US population, that’s difficult to serve, and depending on your point of view, you can be optimistic or pessimistic about that. In other pieces of that, that make me optimistic. There are people in the United States who just don’t want the internet, and they never will. And then there are parts of it that make me pessimistic. There are Native American reservations in the United States that have just, you know, it’s offensive, this terrible service. So, that’s the sort of a technology version of this.

Scott Wallsten:

So, with the pandemic, [inaudible] I guess, ask lots of professors, their views on things, and just in this one case, they asked you whether… the statement was, “The COVID-19 pandemic will lead companies to relocate infrastructure deploys away from dense urban areas,” and you agreed with the confidence of nine. I’m assuming that’s on a scale of one to 10. Now, I guess, assuming…

Dr. Shane Greenstein:

I still do so. 

Scott Wallsten:

Right. So, what does that mean for the pessimistic and optimistic views? Does it mean that we’ll be pushing some of those benefits further out beyond the urban areas?

Dr. Shane Greenstein:

Well, it means one thing for certain, that if you’re a real estate broker in San Francisco or Manhattan, you’re having a really lousy year.

Scott Wallsten:

Right.

Dr. Shane Greenstein:

That’s literally what it means. As [inaudible] to any of the densest, urban areas in the world. It’s not unique to the US. Where people commuted into dense… if you think it through for a moment, “Hi, let’s put myself through a half an hour of misery in the morning, and a half an hour of misery in the late afternoon, so that I can be in the same location with somebody else, doing a job that…

Scott Wallsten:

Somebody else you may not want to be with.

Dr. Shane Greenstein:

That actually you could just as productively do at home. So, it turned out the forced experiment… Look, the pandemic is awful. So, I don’t want to, this is trying to find humor, its gallows humor. Okay, so just make sure, you know, your listeners don’t misinterpret me.

Scott Wallsten:

Everyone knows, you’re a good person.

Dr. Shane Greenstein:

This is gallows humor. As you find, you know, the forced experiment of trying to find, and being forced to find productive time by working at home actually led to the insight that a potted community was not necessarily useful and valuable. And I mean, the data is already in. The amount of remote work in the United States was far larger than what was forecast and far larger than what was typical prior to the pandemic. I mean, the numbers coming in or something like prior to the pandemic, 25% of the US labor force spent some amount of time during the work week doing work at home, and it was typically less than half of their workweek. During the height of the pandemic, it was close to 40% of the workforce spending their entire time at home, and the data that we have so far is that it was more productive for those who got rid of commute time. It got rid of a lot of bothersome activity at work. 

And it was less productive, and let me be very clear, right? It’s not all good. It was less productive for those who had young kids at home. It was less productive for those who also had to deal with issues at home and were unable to deal with things at work that would have raised their productivity.

Scott Wallsten:

Right, and I think lots of women have left the workforce. And lots of…

Dr. Shane Greenstein:

The worst case was, you know, a biologist with young kids at home who was trying to raise mice in the lab. That’s sort of like your worst-case scenario, and there was big variance between small firms and large firms. Large firms around you a much better handle this much better. So, if you had, you know, the, the dumb luck of working for a large firm that happened to invest in the extra things that you needed, that also worked out. 

Having said that, boy, the experiment indicated to a lot of companies that you can survive this way, and that maybe having people come in two or three days a week, rather than five is, you know, all around, everybody wins out. You need less real estate to get more time. You can take care of the things you need to do at work when you’re there, and for the times when you don’t, you can Zoom. It turns out that’s good enough.

Scott Wallsten:

I mean, it is amazing recognizing all of the terrible economic consequences for so many people, but you’ve got these, you know, entire downtowns that are mostly empty, and the economic effects of that, at least measurable economics are so much smaller than you might have expected.

Dr. Shane Greenstein:

Yeah. It’s a little unclear, right. Because we don’t know how much of this is landlords just taking a hit right now and hoping it’s going to come back. 

Scott Wallsten:

That’s true. 

Dr. Shane Greenstein:

Yeah, and then the other thing, the other forecast that’s being made as you know, those buildings are not going to go unoccupied. So, I mean, I lived in Chicago for many years. Absolutely wonderful city, and the downtown is just awesome. But you know, if that real estate isn’t going to be used for offices, people going to live there. I mean, those buildings could, many of those buildings could quite easily be converted to residences. So, it’s, you know, it’s got to be used for something. I got to admit, living in downtown Chicago was so wonderful.

Scott Wallsten:

Actually, yeah. I mean, all of those offices with all those nice windows, they would be beautiful places to live.

Dr. Shane Greenstein:

Yeah. They’re beautiful places to live, but at the moment they’re beautiful places to work.

Scott Wallsten:

Right? Yeah. We’ve talked a lot about counterfactuals and, and you’re in a pretty good place to see the things that are studied. Sarah, we were talking about this earlier. You had some really interesting thoughts, questions about that.

Sarah Oh:

About studying the counterfactual? 

Scott Wallsten:

About what Shane sees at NBER.

Sarah Oh:

Yes. I was curious to hear about that. So, as a Co-Director of the Program on Economics with Digitization at NBER and Economics of AI, you must see a lot of paper proposals coming through your door and you do journal service. So, you’ve probably seen hundreds of the newest research papers for digitization. What kinds of papers would you like to see more of? What’s the quality of research questions do you think in this era of digitization, and do you have any advice for researchers studying these topics?

Dr. Shane Greenstein:

Wow. So, let’s divide the world into consequences for households and consequences for business. Even though those overlap sometimes, like in remote work. Okay. So, the consequences of digitization for business is still an enormous area, and it’s the one area, certainly at a broad level where I wish there was more work, and it’s partially because it’s so difficult, but what are the productivity consequences of digitization? We have accumulating evidence that the digitization influences the frictions, you know, reduces the frictions of transactions, and so that changes the composition of activity people undertake. So, you know, it’s easier to communicate across distances. What we’re doing right now is easier. It’s easier to monitor your own employees. It’s also easier to monitor your partners. But what were the consequences of that? That’s been very hard to document. There’s sort of a little bit of evidence here and there that firms rearrange the location of where they have activity. There used to be forecasts. Scott was alluding to this a little bit earlier, too, that everybody would just go to their cabin in the forest and work from the forest, and then economic activity would spread out from the city center, which did not happen. But during the pandemic, actually, quite a lot of that did happen, and it was very uneven. In other words, it happened to some parts of the economy and not others, and again, to get evidence of that would be terrific. 

Another thing that I don’t see much work on, but we kind of have a sense about, which is the central cities became much more important in the global economy. So, a different way to say it is the forecast about things moving out of New York would have been that would benefit Des Moines. That doesn’t seem to be what we’ve seen. What we saw instead was it benefited London and although maybe, you know, or Tokyo or Singapore. So, it intensified the importance of the major cities and did not seem to help cities that were of a smaller size around North America, and that’s interesting, and I don’t know why. 

Something that’s gotten… I’m just giving you lots of things. We see evidence that there was not an improvement in incomes for middle income and lower incomes from digitization. We see plenty of evidence of improvement in income for those who worked in information intensive markets, intensive firms. So, finance, IT production, some parts of entertainment, especially the best parts of entertainment, you know, the better parts, but we don’t see it in middle-class income. For sure, we don’t see it in a middle-class income, and the question is why? That’s a still a big question. Okay. So that’s on the business side. I was just like, that was just business. 

At a household, you want to divide the world into before the smartphone and after the smartphone. So, we have deployment of broadband in the early part of the new millennium and continuing thereafter, and that’s got this characteristic that households benefit because they can now communicate with all these firms. They can access all these firms that were previously inaccessible. It’s also two way, right? Firms could suddenly access all these households in ways that were not previously possible as well. In the midst of that, I mean, we were only halfway through, that would have been a major change in distribution. Electronic commerce was going to grow as a consequence of that. Entertainment was going to change as consequence of that. In the midst of that, this is why counterfactual is hard here, we got the smartphone, and the smartphone makes that mobile. And again, we know from data already half of the time spent online has moved to mobile devices. So, you know, that’s a different world than we would have had, had we just only had broadband. 

Okay. So having said that, what would I love to see? We don’t see estimates of consumer surplus from the smartphone or for broadband. Another thing we don’t see is the uneven benefits, the different kinds of activities. Changes in use of household time, Scott worked on this some years ago, and I was so happy when we saw it. I put that chapter in, you know, in a book I edited, and I hoped it would be a catalyst to lots and lots of future work on changes in the composition of time at home, and except for work I did, it didn’t happen, right? But I’d love to see…

Scott Wallsten:

I guess we said everything that needed to be said. 

Dr. Shane Greenstein:

Yeah, that’s right. But you know, I watch my kids and the way they allocate their time is so different than the way certainly I allocated it as a child. They spend almost no time on television. It’s almost all on their phones, and that’s really remarkable. They do almost no programmed television. It’s all streaming, if it’s anything, or YouTube. The way they consume sporting events, this is about the only programmed thing my children watch, they are on Twitter simultaneously at the same time as the sporting event, and they’re consuming that as much as they are the sporting event. So, changes in those sorts of things and the how to think about the welfare consequences of that are a pretty big deal, and we know we’ve had a little bit of work on those sorts of issues because it’s shown up in the election. So, when it shows up there, people think about it. It’s actually underneath the surface.

Really, this is a pretty major change in the quality of life in one generation. I always say, “Would my grandparents have recognized what we are doing?” So, at first, I always say the same thing. I always say it was just, you know, like my grandfathers who raised chickens and cleaned the laundry, that their first reaction would be the same one, which is, “You get paid to do what?” That would be their first reaction, and then their second reaction would be, you know, “You can fly across the country. Oh, except you have to wear a mask. What?” And then the third one would be “Holy cow, what do you spend your time at home doing, what in the world?” They wouldn’t recognize it at all. So just to give you an example, and we don’t, it’s funny that it’s taken a while for that to catch up economic research to catch up to that.

Scott Wallsten:

You know what? This is taking a couple of things that you’ve said thinking more broadly. So, you know, the internet was a series of incremental innovations, and the smartphone was a transformational innovation. I mean, it built on lots of incremental things, but there’s one event, but the internet was a series of smaller events. Do we learn anything from that about what the economics of innovation need to look like?

Dr. Shane Greenstein:

Oh, there are two great insights in the smartphone. One is that incremental change in the underlying technology can nonetheless lead to dramatic changes in the experience. That’s actually a well-known theme from the history of technology. So, a whole series of incremental things accumulate to get you a radical change in the experience, and there’s lots and lots of examples of that from the history of technology. The one example that sort of, that we always use in the classroom is computing gets better continuously, being able to use advanced computing to take a picture of your body in a CT scan is an enormous change in medical capabilities, and then it gets continuously better again, and finally, good enough that you can take a picture of your head, and to be able to take a picture of the skull and identify a brain bleed has enormous consequences medically, but it’s of course it’s continuous improvement until you get above at some threshold where it’s sufficiently good, and then there’s been continuous improvement since then, and now it’s amazing. I know people who were saved by that, you know, and that’s amazing. 

Okay. So, coming to the smartphone, there’s a whole series of continuous innovations that are affiliated with the smartphone, including touch and incremental changes in the speed of transmission of data over the carriers. You have got to remember going from 2G to 3G important consequence for the smartphone, and boy, to 4G had really big consequences. Putting more computing on the device has enormous consequences. Again, it was continuous. Memory was getting cheaper and better. Computing was getting cheaper and better, and you could have used a smartphone in 2001 and 2002, which is called the Blackberry. So, there’s that. 

So that’s one, that’s one sort of major observation, other second observation, and I’m going to go back to something I said earlier is the advantage of having multiple points of view in technically advancing markets. So, the smartphone was not a new idea. Microsoft had a smartphone. I mean, I’ve met the guy who was in charge of Windows CE. I mean, you know, he was given an impossible task and told, you know, recreate what happened in the PC, and that was in 2001 when there wasn’t a single firm on the planet that trusted Microsoft, right? So, you think that one through. Poor guy, right, given that assignment, a wonderful person, individual, but given that assignment in 2002. “Hi, I’d like all of you to co-partner and invest with me, even though I work for a firm that all of you felt burned by a few years ago.” So how do you do that? 

Well, the best thing about commercial markets is other firms show up with alternatives. The Blackberry shows up in that environment and we get a whole series of experiments from other firms and then find who, we get the experiment that ends up catalyzing the market, if you were to describe it in advance, nobody in their right mind would have thought this was going to work. 

We’re going to take a CEO on a second, who people hate. That’s very clear. Unloved by his entire ecosystem. So not unlike Microsoft, right as the CEO is not liked at all. Completely despised by everybody in the Valley who’s ever had to deal with him, except the people who work for him, who love him because he has their back, and this CEO has successfully deployed the iPod, if you recall, which was a portable music device and is claiming he’s now going to make a phone call on such a device.

Okay, that sounds crazy. Who in their right mind thinks that’s going to appeal to anybody, and again, if you go and you watch the first announcement, there’s nothing about apps in there. There’s nothing about an ecosystem. It’s all about carrying your music around with you and make a phone call. That’s the entire appeal, but it generates an install base that’s large enough that it generates an app development process, and the firm who’s doing that has lots of experience with the app development process. So, they’re very good at that. In that sense, they’re identical to Microsoft. They’re very, very good at supporting app development, and it catalyzes action from a third from, who turns out also to be very good at supporting app development, which is Google, who has purchased a few years earlier, by the way, Android, and catalyzes development there.

Think about it. If you go back, Nokia had already… Do you remember Symbian? Nokia was trying to support app development. They were terrible at it, right? That wasn’t working. Blackberry was terrible at it, and they weren’t doing it very well. There were other attempts. So that’s what commercial markets are great at, that we get multiple attempts. And of course, Scott’s correct. Then we get, finally, we get a catalytic moment where the apps and the device interact to deploy and get things started, and at least we get two major ecosystems in the world. It’d be kind of nice for competitive purposes if we had a third, but for the moment we still have two.

Scott Wallsten:

We used to have a third competitor in these that they didn’t work. We had a Blackberry and we had Palm and neither of them succeeded. But are there, are there other competitive issues that you’re concerned about?

Dr. Shane Greenstein:

Well, there is an antitrust case being filed right now. 

Scott Wallsten:

That’s true. 

Dr. Shane Greenstein:

It’s a bit hard to get uncontested facts, but you know, if you read the DOJ filing, there’s a number in there that says Google pays Apple somewhere between $8 and $12 billion a year. And that is related to making Google the default search engine on the iPhone, as well as making it the default search engine on other Apple products. According to news reports, approximately $8 billion of that goes just to the iPhone, just to make the Google search engine the default search engine when you buy an iPhone. So, without ever making any choice, the search engine comes up immediately with Siri and comes up immediately with Safari as well as other programs. So, let’s think about that for just a minute. Here, we have two firms that are competing in the smartphone market and a feature of one of those firms is being purchased by the other. It’s being supplied by the other.

These two firms, both are extraordinarily profitable among the most profitable firms in the history of Western capitalism, and one of them is paying the other at least $8 billion a year to become the default search engine. Okay. So, let’s just think about that for just a minute. We have two firms splitting the market, and one of them, Apple, could have been a provider of search engines. They have $8 billion worth of reason not to be. We have two firms ostensibly competing in the smartphone market, and one of them is setting a default that helps the other. So, we now have an incentive not to compete too fiercely because we both are benefiting jointly from keeping our duopoly nice and calm and keeping prices high. We also have a setting in which distribution is very difficult to break into. So, if you are the proverbial from in the garage, you don’t even bother.

Now, if you’re a VC and you see something that’s opportune, you don’t even bother. If you’re a moderately sized firm and you want to do a new feature, you don’t bother. You have a setting in which dynamic competition is not operating because distribution is not unrestricted. It’s been purchased by the leading firm. I don’t understand why this is not a per se violation of antitrust. You know, you have two very lucrative firms making a cash deal, usual instincts. I mean, we all learn these usual instincts are one firm is not allowed to use cash to determine the characteristics of its rival’s product, end of story. This should look extremely suspicious. John Sherman, who ran the antitrust, right? John Sherman, the Sherman antitrust act. He hated these kinds of deals, and there were reasons he hated them because they look so suspicious. 

Another way to think about it. Why is Google spending $8 billion to purchase distribution when they could have been using that $8 billion to make their product better? Isn’t society better off if they spend the 8 billion on making search better?  I’m not saying it’s a bad product. I’m just saying going down, you follow the logic of this kind of deal, at some point, it’s not good for society. Now, I want to be clear. I use Google products. I don’t hold it personally against them, but I want them to earn that. And I also want them to earn it from people whose job it is to review Google products, because I read those. And if I read that they’re not doing a good job on something, I’ll go try something else. So, it should make sense for society to want Google to take that money and devote it to making their product better.

Scott Wallsten:

So, let me play devil’s advocate here. So, Google is paying Apple to be the default search engine and the default search on probably several, several products. And so, I guess there are a couple of questions. One is, you know, I suppose the answer in a trial to that might be well, that just shows how scared Google is of entry that they’re willing to pay that much money to make sure that they are on Apple’s phone. But then also, even if that’s not the case, you’re then jumping from that to saying that it keeps them from competing in the mobile operating system market. And I’m not sure how, I’m not sure that that logic tracks.

Dr. Shane Greenstein:

Yeah, yeah. I hear you. Yeah, let’s do the first one first. This is, you know, an old idea in antitrust, distribution ought to be open and unrestricted.. So, let’s just do that one, and the firm who has the most interest in order to restrict distribution is always the firm misleading. And the best way to avoid competition is to restrict distribution. So, at least I’ll go straight to that one. On that principle alone, it’s enough to say this looks suspicious that you ought to have a much more unrestricted choice given to users at the beginning of their use of the iPhone, and you could imagine an antitrust settlement imposing such a restriction. There’s already precedent for it in Europe. Hesitant, as I am to say, the EU gets anything right.

Scott Wallsten:

Counterfactuals are horrible because we never know what they are, or, you know, what they could be, but in this case, what the counterfactual is Bing. So, if they had entered into an agreement with Microsoft, would that be okay?

Dr. Shane Greenstein:

Yeah. Yeah. I know where you’re going. I love this one! Let’s do that. Let’s even do a better counterfactual. Don’t you remember back when people were complaining that Microsoft had a hundred percent of the browser market, and they ought to be compelled to let in an alternative, and if you looked at the situation at the time, you said there isn’t going to be an alternative because there weren’t any.

Scott Wallsten:

So, this agreement is at least partly responsible for the lack of alternative search engines, I guess?

Dr. Shane Greenstein:

Yeah. I mean, it’s endogenous. The lack of alternatives is endogenous to the difficulty of getting them into the market. And, let’s remember also what happened in browsers. It didn’t happen right away. There was actually an open requirement imposed that Microsoft and I think more importantly, there was embarrassment so that the company became hesitant to be so heavy handed in the future. And what happened over time? Actually, the alternatives did emerge, and there was competition in browsers, and let’s not forget Microsoft didn’t have a very nice browser 15 years ago. And I got to say, actually, I think their browser today is pretty good, but

Scott Wallsten:

It does seem to be pretty good, yeah

Dr. Shane Greenstein:

The competition had an effect on them and isn’t that great? 

Scott Wallsten:

Although, you know, there are elements of how the browser market developed that seemed to support parts of both sides, because on the one hand, everything you said is exactly true, right? And we’ve got innovation and browsers, first we got Firefox and we got Chrome. And then we have lots of projects built on Chromium and, and so on. But we still ended up with another dominant browser. Mabye, that’s what the market is? I don’t know.

Dr. Shane Greenstein:

Although, you know, it’s… I appreciate your point. If you’re going to have dominance though, you want to have I think society benefits from having a constant threat that, that dominance is going to switch between firms, and anything we can do in antitrust to make that threat higher is good for society. I’m a firm believer in that now I… you know, I appreciate your point. Counterfactuals are hard and Google, to be fair about it, is really good at what they do. And, and maybe people would still choose it if given the option to choose. I sort of, I would now too, but I don’t know what I would do two or three years from now, if I had the option, and I’d like to have that option, but I don’t. You know, I actually take offense that that choice was made for me, and I didn’t even consciously know that it was possible to switch. So, that that’s an easy thing law can do for us.

Scott Wallsten:

What about the harder one though? Whether it extends to other, other markets, like the mobile operating system?

Dr. Shane Greenstein:

Yeah. There are a lot of hard problems. I don’t know. Where do you want to start, Scott? That’s a hard problem.

Scott Wallsten:

You brought it up.

Dr. Shane Greenstein:

They compete in the smartphone market. Every decade or so another device seems to show up. I don’t know what it’s going to be, and I don’t know when it’s going to be. But I’d sure like, when it does show up, that not only do these two firms compete for it, but lots of other firms compete for it because they believe that distribution is not locked up. So yeah, that’s a little bit to keep things open in order to keep our options better in the future. So, it’s a future benefit and it’s kind of abstract. So, I could understand resistance to that. But you look at US antitrust. We have more dynamic markets, and that’s for a reason. And one of them is because we, from time to time have many options. 

Scott Wallsten:

Yeah. Right. We mean what we talked about, Blackberry and Palm. And Microsoft did have it… introduced a Windows, a real, you know an actual mobile operating system, not the one that had like the little start button and didn’t have enough memory for you to answer the phone when it rang. But, you know, an actual operating system, and I had a phone that was windows. It was great. I loved it, but it didn’t take off.

Dr. Shane Greenstein:

Not to pick on Microsoft, but, you know, they were in the cloud quite early. It just, wasn’t very good. And now that you know, Amazon has shown them how to do it, to Microsoft’s credit, they’ve gotten their act together and their cloud services are pretty good. And you know what, that’s great. That’s, you know, and that’s the way the internet ought to work. That’s how competition ought to work. And, Google got into the cloud market with some of their services and their services are particularly popular with the tech crowd, and AWS is behind with that crowd. And so, they feel the pressure. That’s good.

Scott Wallsten:

That actually might be a good note, an optimistic note to leave it on. 

Dr. Shane Greenstein:

Optimism is good!

Scott Wallsten:

That was supposed to be an inherent part of technology, right? That we don’t seem to have any more for maybe good reasons, but anyway, it is nice to have an optimistic tone. And so, Shane, thanks so much for being with us. This was a lot of fun.

Dr. Shane Greenstein:

Pleasure. Thank you for having me. It was fun too. It’s always fun talking to you.