Before he became the Adam Smith of Googlenomics, Hal Varian spent decades as an academic economist, writing influential papers, a popular book about the information economy, and several textbooks that are still taught today. So how has his nearly twenty years in the business world affected what he’d write and teach now? Is learning Shephard’s lemma really that important anymore?
Tyler asks Hal these questions and more: why aren’t there more second-priced auctions — or prediction markets? How have the economics of sales changed with the internet? In what ways did his hiring criteria change between academia and business? What could we learn from the sack of Rome? When should economists avoid looking at the literature? How are we always eking out victory in the war on spam? And what are people least likely to understand about Google? Fear not — Hal has an answer for it all.
Listen to the conversation
Read the full transcript
TYLER COWEN: Today, I’m very happy to be here with Hal Varian, who is chief economist at Google and emeritus professor at Berkeley. Hal, welcome.
HAL VARIAN: Thank you. Good to see you.
COWEN: Now, one of your most famous articles — it dates from 1980, and it’s about a theory of the economics of sales. One motive for sales is price discrimination. But what’s the main alternative explanation of why sales happen?
VARIAN: Well, the claim is always overstocking or fashions changing or selling something at a loss or a remainder sale.
COWEN: But even that’s a bit like a price-discrimination explanation. It’s a Coasean durable goods monopolist — you’re just unloading it at a loss.
VARIAN: Yeah, but they’re unloading at a loss because some new models come along that’re superior. They’ll cut the price.
You know, in France, you’re only allowed to have sales twice a year, and they have to be legitimate price discounts. Kind of a remarkable fact I discovered on a visit there.
COWEN: When you say legitimate price discounts, what counts as legitimate?
VARIAN: They had to be charging a higher price before and then actually cut the price from the previous high price.
COWEN: Now that we have the internet, how has the economics of sales changed?
VARIAN: The economics of sales has really found its home on the internet because you look at users: Some people search for the lowest price, and they have all these tools to find it. Some people just buy on an as-needed basis. So you’ve got these two segments of the population that fit the model that I wrote in 1980 really quite nicely.
COWEN: Should we be more or less worried about the economics of price discrimination, given the internet?
VARIAN: We hear two stories, of course. One is, there’ll be this highly personalized pricing that will leave consumers with no surplus whatsoever. That story seems, to me, to be ridiculous because it’s so easy for consumers to conduct this kind of search and try to find a lower price. So consumers have been empowered by this technology in my view.
COWEN: Let’s say I’m searching on the internet. Amazon has a lot of data about me. Say they were completely allowed to price-discriminate. What’s my response to this hypothetical Amazon third-degree perfect price discrimination? What should I do?
VARIAN: Use Google. [laughs]
COWEN: And buy from someone else?
VARIAN: Sure. By the way, this is not an uncommon behavior because lots of people will go to a big retailer — it might be Amazon, it might be Macy’s — and they’ll get a price. Then they do a little search to see if they can improve on that price. It’s a model that’s not at all unusual, to see people describe that kind of behavior as, say, a normal way of online shopping.
COWEN: But the size of that convenience wedge — is that captured by Amazon? Or how much of that surplus do I get? I’m set up for one click on Amazon, right? So I want to use Amazon, all else equal.
VARIAN: Well, how much would you charge to do two clicks? It’s one of these things. Yes, there’s a minor extra cost because the site is trying to make it as easy as possible for you to buy, but at the same time, they can’t really impose any significant cost. So the most they can make is the difference between your valuation of one click, two clicks, whatever.
COWEN: Another topic you’ve written a great deal about: electronic journals. Today, why are so many of them still priced so high? Marginal cost is zero, right?
VARIAN: Right.
COWEN: They would appear to be semi-close substitutes.
VARIAN: And they are semi-close substitutes because, as you know, you’ll often find copies of published papers or preprints or something available online. The end users — the faculty members or the students — they can usually get what they want, so they don’t protest these high prices much.
It’s the libraries that end up paying. And the librarians, generally, say, “We want to provide useful information to our constituency.” So they’re not highly motivated to resist these high prices either.
COWEN: But I can use Google to look for substitutes, right? To bring it back to the Amazon comparison.
VARIAN: Of course.
COWEN: But yeah, journals will still charge, say, a few thousand dollars a year for a subscription.
VARIAN: But not to you. It’s to the library or the university you’re at, is the basic answer. So there’s a little bit of a division of a problem with incentives here in terms of getting you to face the true cost of that journal access.
COWEN: Why don’t libraries rebel more?
VARIAN: They do to some degree. And you’ll see situations, especially at smaller places, where they will arrange an interlibrary loan or some system that they can ride on the larger organizations’ coattails. And that’s not at all uncommon.
COWEN: Why don’t pirate copies now break down the equilibrium? I could go to Sci-Hub, right? Which is black or gray market, but it’s there. No one will stop me.
VARIAN: Well, again, the main users are going to be people at universities. The universities already have a license for these things. So yes, you might go to Sci-Hub if you can’t find it elsewhere, but just like you said before, the path of least resistance is just search over my library connections to see if we have access to this material.
COWEN: But if someone says that, well, the equilibrium is a kind of market segmentation, where the ready searchers get it very cheaply or for free. The other people are elastic. If Amazon is perfectly price-discriminating against us, do we end up with the same situation? Is it like the journal articles? Or is it, somehow, we all search and Amazon’s price discrimination breaks down altogether?
VARIAN: Well, I would say — if we take the journal example again — as we all know, where we’re going to get the price reductions is competition. And you’ve probably seen that in the last couple of years, the American Economics Association has offered four new journals. You’ve got a bigger set of publishing outcomes. There’s a high level of prestige attached to these journals, and they’re very reasonably priced.
I think we are seeing entry of that sort, not only in economics, but in other scholarly disciplines as well.
COWEN: Why are textbooks still priced so high? Not all textbooks, but many.
VARIAN: They are priced remarkably high, and it’s a situation where I really would like to see lower prices because, obviously, there’s a durable goods monopoly problem there. As you have more and more competition from previous editions, each of the new editions has to differ markedly from the old edition to support the pricing model. But that’s getting harder and harder to do.
In fact, a friend of mine once told me, “Having a successful textbook is like being married to a very wealthy person you don’t like much anymore.”
COWEN: Why doesn’t the world use more second-price auctions?
VARIAN: Well, that’s quite an interesting question. Especially now, you’re probably seeing this note that Google was switching to a first-price auction on display ads.
COWEN: Yes.
VARIAN: And the reason they did that is because they wanted to put all of the ad sellers on a level playing field. There was concern — I think unfounded concern, but I will say there’s definitely concern about having the last look at the price. If you are going through a series of auctions, as you often are, then the person who gets to look at the last price could say, “Well, do I want to beat that or not?”
We all know from the revenue equivalence theorem that you would expect a continuous first-price auction to end up about the same place as a one-time second-price auction. So we’ll have a nice chance to see how well that theory works in this context.
COWEN: Why don’t we see more descending-price auctions? Just start with a high price. Keep on lowering it. When someone says, “It’s mine,” it’s theirs.
VARIAN: Or, even better, the first person to push the button gets the item, but he has to pay the price when the second person clicks the button. So you’ve got a combination of a sealed bid and a descending auction.
That’s a very good question. Of course, we do see it. It still happens in the Netherlands at the flower auction. If you are passing through Schiphol one of these days, it’s a five-minute taxi drive. You can go see the market in all of its glory.
But I don’t really know. I think you need to have this equipment, this infrastructure that will allow you to do it. That used to be kind of hard, but now, with a mobile phone, it just comes down to being an app. So, why? I don’t know.
COWEN: If I buy something at Sotheby’s, there’s the price I pay for winning the auction. And then I pay a 15 percent buyer’s premium. Why does that make sense? Why does that persist? Is that a behavioral quirk? That people think they’re getting it for cheap, and then the 15 percent is tacked on, and they’re fooled?
VARIAN: Well, remember, Sotheby’s is offering some services. They’re offering you some guarantees about whether this is an original. They’re offering you certain amenities and so on. Now, whether that’s worth a 15 percent fee, that’s up to you to decide.
COWEN: But why is that a percentage fee? There’s other ways you could charge people for their services, right?
VARIAN: Sure. I think it’s two big firms doing this kind of antiques, fine arts auctions and so on. They’ve established these patterns of behavior, and I’m sure that we will see challenges to that in the future because it’s so easy to experiment with auction design now.
COWEN: Do Sotheby’s and Christie’s have anything to learn from Google about auctions? Or vice versa?
VARIAN: Well, remember, they normally use this ascending-bid English auction, and as you know, the second-bid auction is equivalent under certain circumstances. But there is some degree of excitement, irrationality, madness of crowds — whatever you want to say — that people get caught up in the bidding. And they probably do make higher revenue from the descending-bid auction than from the theoretically equivalent Vickrey auction.
COWEN: How much do you believe in winner’s curse?
VARIAN: Winner’s curse would be most relevant when there is some asymmetry of information. I think, from what I could tell looking at the experimental literature, that naïve players do have a winner’s curse, but they learn over time to shade their bids in a way that gives them a better outcome. So there’s a learning behavior going on there.
COWEN: But if there aren’t many auctions where everyone is experienced, then, on average, we see winner’s curse?
VARIAN: When I started out at the University of Michigan and wanted to furnish my first house, I would go to auctions. One of the rules of thumb that I developed was, you never go to the auction that has the best stuff because everybody’s there. You go to the auction that’s sort of second best because then you have a better chance of beating the dealer.
When I started out at the University of Michigan and wanted to furnish my first house, I would go to auctions. One of the rules of thumb that I developed was, you never go to the auction that has the best stuff because everybody’s there. You go to the auction that’s sort of second best because then you have a better chance of beating the dealer.
So there’s a lot of — I don’t know — common-sense sort of strategies that people use in this, and I think there are inefficiencies that could be exploited, at least at these small-scale auctions with nonprofessional buyers.
COWEN: Why doesn’t business use more prediction markets? They would seem to make sense, right? Bet on ideas. Aggregate information. We’ve all read Hayek.
VARIAN: Right. And we had a prediction market. I’ll tell you the problem with it. The problem is, the things that we really wanted to get a probability assessment on were things that were so sensitive that we thought we would violate the SEC rules on insider knowledge because, if a small group of people knows about some acquisition or something like that, there is a secret among this small group.
You might like to have a probability assessment of whether that would go through. But then, anybody who looks at the auction is now an insider. So there’s a problem in you have to find things that (a) are of interest to the company but (b) do not reveal financially critical information. That’s not so easy to do.
COWEN: But there are plenty of times when insider trading is either illegal or not enforced. Plenty of countries where it’s been legal, and there we don’t see many prediction markets in companies, if any. So it seems like it ought to have to be some more general explanation, or no?
VARIAN: Well, I’m just referring to our particular case. There was another example at the same time: Ford was running a market, and Ford would have futures markets on the price of gasoline, which was very relevant to them. It was an external price and so on. And it extended beyond the usual futures market.
That’s the other thing. You’re not going to get anywhere if you’re just duplicating a market that already exists. You have to add something to it to make it attractive to insiders.
So we ran a number of cases internally. We found some interesting behavior. There’s an article by Bo Cowgill on our experience with this auction. But ultimately, we ran into this problem that I described. The most valuable predictions would be the most sensitive predictions, and you didn’t want to do that in public.
COWEN: If you were a young entrepreneur, what industry would you look to improve upon using mechanism design tools?
VARIAN: Health, of course.
COWEN: Health. But how so? How can we improve health?
VARIAN: There are all sorts of screwed-up incentives. If you start thinking about it from a mechanism design point of view, you can imagine situations where you could improve those incentives by various kinds of marketplaces, market mechanisms.
Now, I’m not an expert in health. I said health because we’re living in an aging population. There’s going to be all sorts of increased expenditures in this area. There’s going to be greater calls for more efficiency and cost reduction. So using a mechanism design tool might really help in some of these cases.
COWEN: The idea of using mechanism design to improve voting — I don’t mean voting in national presidential elections, but ways you could do it noncontroversially: Demand revelation mechanisms. Glen Weyl’s quadratic voting. What do you think of those ideas? Can we actually use them? They’re hardly ever used. They seem to make good economic sense. They haven’t evolved.
VARIAN: Right, but we have seen things like instant-runoff voting, which Australia has been doing for a hundred years. We’ve got now Maine doing the same mechanism. Eric Maskin talked to me a few days ago. He’s working on an issue to get Massachusetts to adopt this kind of system, and I think he puts forth some pretty compelling reasons.
COWEN: But why are we so reluctant to incorporate preference intensity? Again, it doesn’t have to be a national election. It could be the local bridge club. You could have people do Groves-Ledyard-Clarke. They could do Tullock. They could do many other schemes. Never seems to evolve other than a bunch of nerds doing it for kicks or research purposes.
VARIAN: This question of intensity — there is one margin, namely, people who feel very strongly about a certain candidate or certain issues, devote their own time, energy, volunteerism. That’s a measure of intensity there. If I can recruit another voter to my side, then it’s like I get two votes, my own vote plus the person I recruited. So, there’s a little intensity going on there as well.
COWEN: Why do people trade so much in financial markets? It doesn’t seem Bayesian rational, right? “Oh, you want to trade with me. I’ll take that off your back.” Yet trade volume is massive.
VARIAN: Well, people say it’s massive, but actually, it’s usually a relatively small fraction of the total amount of the asset that’s held. Even on a very heavy trading day, that might be a few percent of the total volume of the assets. So it’s not so big.
COWEN: But you might think you only trade when your kids are about to go college, and you need to write a big check.
VARIAN: Yeah, I agree with your point that there is more trading than there should be by any reasonable model. Part of it is because people really do have differences of opinion, and they’re not fully Bayesian, so they may not find the other person’s opinion credible. We don’t really get the agreeing to disagree or, I guess, the converse. We don’t get this pushed into the model where you’ve got full agreement.
I actually did some work in this area several years ago, and it really came down to people do have a different model. We can’t agree on the model. If we don’t agree on the model, then we won’t get uniformity.
COWEN: And we’re also not meta-rational across models?
VARIAN: Right. We can keep going turtles all the way down.
COWEN: I don’t trade, by the way, if you’re curious to know.
VARIAN: Well, that’s good. I would say, yeah, why trade? You shouldn’t be trading. We know that just as an empirical fact.
COWEN: Here’s a piece of yours that’s not commonly associated with you: “Are There Psychological Barriers in the Dow Jones Index?” Once it hits a certain level, then it’s broken through, and you can predict future returns. True or false?
VARIAN: It’s very interesting you raise that question. What happened was, back in the . . . Oh gosh, when would this be? In the ’70s sometime, the Dow Jones approached 1,000 and then backed off. And they did it 23 times, I believe. I don’t remember the exact number, but it was multiple times.
People were talking about the barrier or the $1,000 barrier, et cetera. And I said, “Hmm, I wonder what shows up historically.” And it turns out at this time, Dow Jones had just published a book of all of its prices going way back to when it was founded. We did a little analysis of that to see if there were barriers — so, if century marks was at a barrier, in the sense — did the stock price behave in some unusual way?
This was not a theory or a hypothesis. We were just looking at the data to see what it said. What we found out was that the price increases accelerated as they approached the century marks in most periods.
COWEN: And that’s true even after people know that?
VARIAN: Well, it’s one of these things. Of course, the index is totally arbitrary, and I would suspect, as the Dow Jones has perhaps receded in importance compared to S&P or other indices, probably this pattern became dramatically weakened, or it was eliminated altogether.
COWEN: Now, as you know, some of your most frequently cited pieces are on the economics of envy, and these are from the mid-1970s. Now, you’ve lived for a while in Silicon Valley. How have your thoughts on envy changed since you wrote those articles?
VARIAN: I think what’s interesting is, there are lots of cases where we’re seeking some sort of fair division. There is something that we agree is equitable, et cetera.
If you look at that theory of fairness that I worked on — actually my thesis was on that topic — everything seemed to work out nicely if you just have a fixed pot of goods, and you wanted to divide them in a fair way. For example, an estate. There were a number of inheritors to an estate, and you wanted to make it both equitable and also efficient. So, that problem we can solve.
The difficulty came when you brought production into it because people could contribute differently — different amounts to the production. Then, how much would they be compensated for their contributions? And there, there seems to be a great variety of views in the world. Different people seem to have different opinions on that, and it doesn’t work out as cleanly as the first story.
COWEN: Does the typical American envy more the billionaire or the next-door neighbor?
VARIAN: I think the next-door neighbor. It’s interesting, the billionaires are like our instance of royalty. The people want to see what they’re doing and where they’re going out and how they dress and all those kind of stuff. But I don’t think it’s actually envy.
By the way, I know a few billionaires, and there’s a lot of cost to being a billionaire in the sense that you can’t go out in public. Maybe you need bodyguards. Doing a trip here and there is a major undertaking because of the people that have to be informed. It’s much better to be a half a billionaire, I think, than to be a billionaire.
COWEN: Do the billionaires envy each other more than poorer people envy billionaires?
VARIAN: There seems to be something of a pecking order there. It depends. Not so much in tech, I would say, but maybe in finance and Wall Street guys, I think. It’s a motivation that’s different than we see on the West Coast.
COWEN: A few questions about education. You’ve written the most frequently used graduate economics text for many years. Do we still need to teach Shephard’s lemma and duality in first year PhD micro?
VARIAN: Absolutely.
COWEN: Why?
VARIAN: When I thought about what should go into that book, I thought about, what do you need to understand the journals? Take something like the American Economic Review. You’ll thumb through the American Economic Review and come up with a list of topics. The two things you mentioned — they’re still mentioned in the research literature. So that’s what you need to know to read the journals, and that’s what the textbook is supposed to teach you.
COWEN: But if you look, say, at the AER today, so many pieces seem based on stand-alone datasets. There’s much less of a model. It’s mostly not consumer expenditure theory. And if they don’t know Shephard’s lemma, are they really at sea?
VARIAN: Well, for example, there is a big debate going on now about how margins have changed over the years. You may have followed as well.
COWEN: Sure.
VARIAN: But it comes right from manipulating first-order conditions, throwing in a little bit of duality. So understanding how that model works is critical to understanding that particular literature. Now, it may be that you know where to look it up, and that may be good enough for some of these things, but it’s just the nature of a textbook. It has to prepare you for actually performing, for actually doing and reading and understanding what’s going on in the economics literature.
COWEN: You’ve now worked at Google almost 20 years. Is that correct?
VARIAN: That’s correct.
COWEN: If you were writing the text again just from scratch with almost 20 years of knowledge at Google — practical knowledge — what would you change?
VARIAN: I would say I have added some of the things that I learned at Google to my undergraduate textbook, which is more frequently revised. There’s a whole section on auctions, the kinds of things we were talking about before, what the ad auctions look like. A lot of strategic considerations showing the student that, “Hey, this stuff is really relevant in practice. It’s used to trade billions of dollars’ worth of assets using these algorithms.”
So, definitely that work has gone into the textbooks. The one thing I would say is that in the textbook — both the undergraduate and the graduate textbook — there’s more . . . what should I say? Exactitude than might be necessary in real life. There’s this line that no battle plan has ever survived encounter with the enemy. And when you look at how decisions are made in organizations, they’re often a lot messier than they are in the textbooks.
We might think of the textbooks as kind of an ideal case, and what we want to allow is that there could be departures from that ideal case. We want to look at robust issues. So I would work on that. And also, by the way, I added a chapter in the undergraduate text on behavioral economics, which I think is quite important in this respect.
COWEN: Has working in the private sector made you more or less sympathetic to behavioral economics?
VARIAN: I would say more.
COWEN: And which behavioral distortions, if one would call them that, do you see are the important ones?
VARIAN: Well, you see a lot of things going on with the consumer’s behavior in terms of saving, in terms of self-control, in terms of just understanding a certain amount of complexity. The usual stuff shows up in behavior, and when you’re doing something like pricing, you want to think about the textbook view, and you want to think about a behavioral view, and maybe what you’re going to end up with is a blend of the two.
COWEN: I see the endowment effect a great deal in worker behavior.
VARIAN: Yes.
COWEN: People think they have intrinsic rights to something they’ve been doing when they actually don’t.
VARIAN: Yes.
COWEN: Has the economics profession moved too far away from models? Some people say there’s been a dearth of new and interesting theories since the mid ’90s. True or false?
VARIAN: What’s happened since the mid ’90s is, we’ve seen this great flourishing of empirical work with a lot of the credibility revolution in terms of coming up with better econometric models.
I will say, a lot of the work that I’ve done at Google has been in the econometrics area. That was one of my fields as a graduate student. And I’ve published articles in Journal of Econometrics over time, and just getting that — being able to develop coherent statistical reporting within an organization is critically important.
COWEN: And you’ve published in your academic career some articles that — at least to some people — would count as, I wouldn’t say obscure, but highly theoretical. You have a 1982 econometric piece, “The Nonparametric Approach to Demand Analysis.” Did doing that kind of work also help you at Google?
VARIAN: Believe or not, that very paper was the inspiration for the model I constructed of Google’s ad auction. Because in the equilibrium conditions, basically, I reveal preference conditions that I would rather be in the position I am in and paying the price I’m paying than to be in some other positions. So there’s a little inequality there. You manipulate those inequalities, very much like the paper you describe, and here you get this nice formula for what the price should look like.
COWEN: But then you must think we’re not doing enough theory today. Or do you think it’s simply exhausted for a while?
VARIAN: Well, one area of theory that I’ve found very exciting is algorithmic mechanism design. With algorithmic mechanism design, it’s a combination of computer science and economics.
The idea is, you take the economic model, and you bring in computational costs, or show me an algorithm that actually solves that maximization problem. Then on the other side, the computer side, you build incentives into the algorithms. So if multiple people are using, let’s say, some communications protocol, you want them all to have the right incentives to have the efficient use of that protocol.
So that’s a case where it really has very strong real-world applications to doing this — everything from telecommunications to AdWords auctions.
COWEN: You once wrote as advice to graduate students, “Don’t look at the literature too soon.” Is that still true?
VARIAN: Yes.
COWEN: And why not?
VARIAN: Because if you look at the literature, you’ll see this completely worked-out problem, and you’ll be captured by that person’s viewpoint. Whereas, if you flounder around a little bit yourself, who knows? You might come across a completely different phenomenon. Now, you do have to look at the literature. I want to emphasize that. But it’s a good idea to wrestle with a problem a little bit on your own before you adopt the standard viewpoint.
COWEN: I’m sure you’ve been on hiring committees at Berkeley and, earlier, Michigan, hiring economists. At Google, of course, you’ve hired economists to work in your group. You’re always looking for smart, hardworking people with good values. But what’s the difference in how you choose the best applicants in those two settings?
VARIAN: When you are hiring a faculty member, then what happens is, they’re all coming in as assistant professors. But when you’re hiring at a business, you may have senior people, junior people, experienced people, fresh-out-of-school people, and so on.
There’s a role that you’ve imagined for them, and part of the issue is to match up the applicant with that role. You might have a really outstanding candidate, but they don’t fit the particular position that you’re recruiting for at that time.
COWEN: But the fresh-out people — what would make someone good for Google but not so great for Berkeley?
VARIAN: Well, it’s a question, if you look at what would make them good for Google, they should have better-than-average computer skills because they’re going to have to deal with the infrastructure we have there, which is a pretty powerful infrastructure but also quite complex. And that would be a strong requirement. Not only having the skills in, let’s say, statistical analysis, but also the skills in terms of computer operation.
COWEN: In the United States, how big a problem is internet congestion today?
VARIAN: Today, I would say, in the US, it’s pretty good because we’ve got quite a bit of fiber out there. When we move to the spectrum, then I think there are issues where we’re going to be encountering more and more congestion, and 5G is hopefully going to come along and rescue us there.
COWEN: How will 5G change my world?
VARIAN: Basically, you should think of 5G as Wi-Fi everywhere so that you’ve got a high-speed communication without having to go through any sort of special operations.
COWEN: But will it save me seven seconds a week, or will it deliver some new and exciting product that I haven’t thought of yet?
VARIAN: When you look at technologies like autonomous vehicles and things like that, they’re dealing with vast amounts of information. It’s often stored and manipulated locally, but sometimes it needs to be shared. Doing that kind of sharing will be easier if you have high-bandwidth 5G technology. But realistically speaking, for most of what you’re going to be doing, it will just save you a small amount of time.
COWEN: Why isn’t there more spam? It’s striking to me. I’ve always had spam, but it’s never been so much that I felt entirely defeated by it. It stayed within a relatively narrow range.
If I think about it theoretically, I wouldn’t expect that to be the equilibrium. I would think of it as an arms race, where for a long time it might be no spam, and then for a year, it’d be spam so awful, I almost couldn’t use the internet. But it stayed in this narrow band. How do we think about that? Or how do we model it?
VARIAN: There’s a nice article in the Journal of Economic Perspectives on “The Economics of Spam.” David Reiley was one of the co-authors. There were a couple other authors as well. And they had something of an opposite impression. They said there was a huge amount of spam. It was quite costly, and it was a major externality that we’ve never really been able to control effectively.
Now, companies like Google and other email providers had developed algorithms that identify spam quite well. And normally, it’s not really a huge annoyance to people because we’ve been able to do the spam recognition technologies.
COWEN: But why is it you always win the arms race? That, to me, is what’s surprising. There’s never been a three-month period where you’ve lost the arms race. And it’s not just Google. George Mason has a Microsoft-based system. It’s not quite as good as Google, arguably, but it works fine for spam. I waste 30 seconds a week fighting spam.
VARIAN: Yeah, I have my private account, and I have my Google account. My Google account — basically there’s no spam on it. My private account — my email address is here and there, and so I do tend to get more spam. And they’re all those mailing lists that somehow you got on and now find they’re not that useful. That shows up in my private inbox too. I don’t know a good answer to your question. I think, as long as it’s under control from the users’ point of view, then we don’t see a big uprising.
By the way, we do see, of course, a lot of resistance to robocalls, which are much more annoying than spam in my experience. And likely there’ll be some regulations trying to control this.
COWEN: Well, there’s been regulation, but it hasn’t worked, right?
VARIAN: Yeah.
COWEN: I’m on the Do Not Call list. I get seven or eight robocalls a day. If I’m at home, maybe more.
VARIAN: So, there’s Nomorobo. Have you seen this? This was a contest that the FTC sponsored to come up with the best system for handling these robocalls. In fact, I put it on our phones, and it seems to work pretty well. So it’s worth a try.
COWEN: How should we improve the underlying governance and architecture of the internet?
VARIAN: Well, this is a very controversial issue, of course. One side says, “The US created the internet. We’ve been overseeing it, managing it in a way that’s been valuable for the entire world.” But other countries say, “No. No. No. We should have more control than we have now.”
And this goes back and forth over time. We’ve been seeing these examples of countries that really have cut off access to the outside world except under very specific conditions. I think it’s quite clear there’s going to be some fragmentation of the internet in the next several years.
COWEN: How many fragments will it be? Will it be two or three or more? The Saudis, arguably, are another fragment on the way.
VARIAN: Yeah. I’m thinking of a half dozen. Something like that. And we can name some examples but —
COWEN: But we all know who they are.
VARIAN: — we don’t have to. We all know who they are, yes.
COWEN: Will there be a killer app, finally, for blockchain other than gray- and black-market transactions?
VARIAN: Yeah. I would like to separate the blockchain from just cryptographic protocols in general. There’s a huge demand for various kinds of cryptography.
Blockchain seems to be, by its nature, relatively inefficient. As an economist, I don’t like this proof of work that this is. I don’t like the fact that there’s one version of the blockchain that has to keep being updated. I don’t like the fact that it’s so slow. There are lots of things that you could fix, and I expect to see them fixed in the future, but I would say, crypto in general — big deal. Blockchain — not so much.
COWEN: When you say crypto — big deal, do you mean cryptocurrency?
VARIAN: No, I just mean things like —
COWEN: It’s cryptography.
VARIAN: — No, I’m thinking about cryptography in general.
Think about deepfakes. How do you deal with deepfakes? You have to have cryptographically signed collections of bits so we can be assured that they really are created by Tyler Cowen or Hal Varian. That is almost certainly going to evolve. And then there’ll be other ways that you can do proof of identity, and that kind of thing, that are more secure, more efficient than what we can do now.
COWEN: Bitcoin is now at $6,000 again, as you may know. As an economist, what’s the most useful model for thinking about how that price is determined? One can say supply and demand, but one needs more than that, right? It’s sort of a currency of some kind that, arguably, could be worth zero.
VARIAN: I would say capital controls is what’s going on. You have countries around the world where capital controls have been biting, and there are several places where people are very anxious about getting their money out of the country because of political instability or other things. I think that is really what’s driving the bitcoin price changes recently.
COWEN: So if every country were Denmark, it would be worth much less?
VARIAN: I believe so.
COWEN: Your book with Cal Shapiro, Information Rules — it came out in 1999. That’s now 20 years ago. What in that book would you most want to update or revise, other than just adding material? What have you changed your mind about?
VARIAN: Ah, that’s a good question. One of the interesting things is, we used, as one of the prime examples of indirect network effects, was operating systems, right? There was a natural tendency to move towards a single operating system, we argued. And I think there is a tendency to do that. But nowadays, we use, daily, five different operating systems easily. There’s iOS. There’s macOS. There’s Linux. There’s Windows. There’s Chrome OS, and there’s Android.
So here are all these operating systems. They all seem to work pretty well together. Not as well as they would if it was designed from scratch, but people have made ways for these systems to relate to each other. The user interface has become standardized. So actually, the importance of the operating system has retreated into the background compared to what it looked like 20 years ago.
COWEN: Is WeChat better than anything the West uses to organize its internet?
VARIAN: I don’t really know the answer to that question because I don’t use WeChat, but it is —
COWEN: But that’s endogenous, right?
VARIAN: Yeah, sure. And what’s happened is, of course, it evolved in a very specific environment, with a different regulatory structure and things like that. So you were able to get this kind of interoperation in a way that would be very hard to do in our economy, our country.
COWEN: Here’s a science fiction question: Do you ever think about a zero-marginal-cost society? Say there’s a universal replicator with super cheap energy, and in essence, all wealth is held in the form of IP. Would that just be crazy? Or could it work somehow?
VARIAN: Well, there’s this wonderful book, a favorite of mine, called Midas World. I don’t know if you’ve ever read it, but I would recommend it. I think it’s out of print now. But it’s about a world where, basically, robots have become so efficient and produce so much stuff that people are assigned consumption quotas. You have to consume a certain number of shoes per month. You have to consume a certain amount of food and a certain amount of this.
And the poorer you are, the bigger your consumption quota is, which I think is nice, obviously satirical. But it’s a very entertaining book because they work through this premise of how this all goes. I was about to tell you how it all worked out, but I’m afraid it will be a spoiler. So I’ll leave that for your discovery later on.
COWEN: Which feature of the 19th century do you think about or admire most often?
VARIAN: The 19th century, yes. Well, I have to say I read a bit in that area. There is a wonderful book called From the American System to Mass Production —
COWEN: Yes, great book.
VARIAN: — which describes this development of interchangeable parts, and, of course, the explosion of innovation that took place in the 19th century was truly remarkable. I’ll tell you, an expert — my grandfather, who was a farmer in Ohio — told me one time, “I was born when people got around on horse and buggies, and I lived to watch men walk on the moon.”
And that was 70 years. It’s amazing what happened during that period. Now, I grant you, that’s the 20th century, but we also saw huge technological progress preceding that in the 19th century.
COWEN: Let’s say you could be sent to the past, protected against disease, with whatever languages you would need. Any previous historical era, and you would live there for six months and observe and learn. Where would you choose and why?
VARIAN: By the way, I have to tell you that’s a premise of another science fiction book.
COWEN: Of course.
VARIAN: That’s a great one I like very much. Great favorite of mine. That is Lest Darkness Fall.
COWEN: Several books, yeah.
VARIAN: Yeah. The American tourist is sent back to Rome around 400 AD, and he takes as his mission the prevention of the sack of Rome and the, basically, destruction of civilization for a thousand years.
What period would I like to see? I guess I would be interested in that period because that would be a double goal.
COWEN: Ancient Rome.
VARIAN: Yeah.
COWEN: The Republic. The collapse of Rome.
VARIAN: Yeah, the collapse of Rome. And, again, this idea of what would you do? How would you acquire resources? How could you put together a system that could prevent that collapse? It’s a clever story. I’ve thought about it many times, actually.
COWEN: I think I would pick pre-conquest Mexico because we have so few written records. We know a fair amount about Rome, obviously.
VARIAN: Yeah.
COWEN: And much less about the Americas before the Europeans came.
VARIAN: Yeah.
COWEN: Do we need to worry more about algorithmic price collusion, not done by humans, possibly not intended as collusion in any way, but nonetheless, in a modeled sense, being quite a bit like collusion? And what should the law do?
VARIAN: Well, you have these examples like rapid-response equilibrium. We have built in this assumption that the consumers move faster than the firms in our textbook models. What happens is, you see two gas stations competing with each other. One gas station cuts its price. All the consumers flow to the lower-price station, and we get this great competition.
But it could be a different model. It could be a situation where the gas stations move more rapidly than the consumers, and one cuts its price. The other immediately matches the price, so there’s been no benefit from price cutting, and the consumers haven’t had a chance to respond. There, we end up with an equilibrium model where you can sustain a super competitive price.
I think there are situations where you can have algorithmic price discrimination. You could think about the —
COWEN: But not just collusion.
VARIAN: Collusion. Sorry.
COWEN: Say I build a smart AI. I just tell it to maximize profits. I don’t intend anything illegal. And then it, in essence, colludes with the other smart AIs, and they keep prices high, restrict output. And I’m just there as an owner. I don’t even know this is happening.
VARIAN: Right, but again, then there’s somebody that’s outside of that little duopoly, and there’s a triopoly and a quadropoly and so on and so on. It gets harder and harder to coordinate all of those activities, especially if you’ve got an agent on the consumer side who’s looking for the lowest price all the time.
So you’ve got to think of these markets where you may have active price setting on the viewpoint of the AIs, but you’ve also got active search and active choice on the part of the consumers. I think it’s an interesting area for research, for sure.
COWEN: Now, you, of course, started off as an academic, an economist. How did you acquire or learn the kind of managerial intelligence required to succeed at Google?
VARIAN: [laughs] I’m not sure if I’ve ever learned the managerial intelligence there. But because Google really started out as an academic spin-off itself — as you know, the Google search engine page rank came out of a National Science Foundation digital libraries grant.
In fact, when I was at Berkeley, I worked on that digital libraries initiative. It spawned three search engines: Inktomi, Lycos, and Google. There was a lot of interest in that topic. Google — when I joined when they were about 300 people, I don’t know that there was a whole lot of management intelligence. Larry—
COWEN: Something went right.
VARIAN: Well, Eric Schmidt came. He described his job as adult supervision, and there is some truth to that. But I will say that you just had a very creative group of people who were doing very interesting work. What was exciting about it was, the intellectual environment there was so thrilling.
COWEN: It’s true that at some point, Brin and Page offered to sell the page rank algorithm to Yahoo for a million dollars?
VARIAN: I think it was Excite. I’ve heard the story about Yahoo, but the most definitive story I’ve seen, I think, was Excite. And the claim in this source was that they wanted to sell their algorithm. They wanted to get back to graduate school and have the life of a graduate student plus $1 million. But the trouble is, they couldn’t really reach an agreement with Excite. Not so much on the price, but on what kind of technology would be used in the search engine.
COWEN: And who showed them the algorithm was actually so valuable?
VARIAN: Who showed them that the algorithm was valuable?
COWEN: Yes.
VARIAN: Back in those days, when you looked at the results from existing search engines plus the Google page rank, it was immediately apparent that the Google system was significantly better than the competition.
COWEN: Google has done many things that economists have advised or recommended. How about the design of the accounting system? Have economists had any input into that? Or are we just worthless when it comes to accounting?
VARIAN: Well, which accounting?
COWEN: Just keeping the books.
VARIAN: Oh, keeping the books. Yeah.
COWEN: That’s just ruled by regulation and law, and economists are useless?
VARIAN: Well, at one point, there was some discussion about building our own accounting system. Finally, that was abandoned because one of the things that accounting systems do — the commercial accounting systems — is they keep being updated based on how the law or the accounting practices or regulations and so on evolve. So the system itself, as a computer system, is not really very complex. It’s just keeping up with a regulatory environment that’s provided by the accounting software provider.
COWEN: Say you were deregulated or regulated more flexibly. Would there be a lot of mechanism design in accounting systems? Or do you think it would look more or less like how accounting is done today?
VARIAN: Well, the trouble is, with accounting, there’re always these gray areas. What’s a good and what’s a service? Or things like that, where there could be a blending from one into another. There is considerable debate about how you should do this, and then there’ll be some kind of reconciliation where people agree, “Yes. That’s a reasonable procedure.”
But I don’t think there is ground truth out there anywhere. I think there are lots of choices that are made to resolve some particular situation that’s come up.
COWEN: Let’s say I’ve read all the usual magazine articles about Alphabet/Google. What’s the thing I’m least likely to know about its culture that matters or is interesting?
VARIAN: Hmm, what are you least likely to know about the culture?
COWEN: Everyone knows about the playrooms, right?
VARIAN: Yeah.
COWEN: Everyone knows about the interview questions.
VARIAN: Yeah.
COWEN: What am I least likely to know?
VARIAN: Boy, that’s a good question. Everybody knows about the mini kitchens. One thing that’s happened in the cafeterias — everybody knows about the free food, but one of the recent complaints was that there was too much kale being served.
[laughter]
COWEN: Complaints from employees?
VARIAN: Complaints from employees, yes, about there’s a bit of paternalism that goes on. They want to give you snacks that are good for you and food that’s healthy. Well, that’s great, but it can be imposed with a light hand or a heavy hand. Some people have argued maybe it’s heavier than necessary.
COWEN: Is there a way to price privacy as a product or feature? I get how it works now. It doesn’t personally bother me very much, but as you know, this is an area of controversy.
VARIAN: I think it’s certainly possible to offer choice, and if you, for example, don’t like the default settings on Google for privacy, you can go in and change those default settings.
Let’s take your interests, for example. There’s something on interest-based advertising. You can go look at your interests. You can edit them. You can delete them. You can add things. You can subtract things. You can opt out entirely, and that capability is completely under your control.
COWEN: But it’s a lot of binary choices. There aren’t markets there, right?
VARIAN: Right. There aren’t markets.
COWEN: And that seems strange. You could have some kind of mechanism design in privacy features. We hardly ever see it with any company. I don’t just mean Google.
VARIAN: Well, we have seen it in the sense that you’ll find some vector of attack that hadn’t been thought of, and then there’ll be a way to control that vector of attack.
For example, a few years ago, there was this question about whether government agencies were tapping into various data providers, but it turns out they were tapping into the telecom infrastructure. And once that was discovered, all of the communication among data centers became encrypted. That was the response to that particular attack. There’s going to continue to be things of that sort.
COWEN: Now, users seem to like them both, but if I just look at the critics, why does it seem to me that Facebook is more hated than Google?
VARIAN: Well, you know, I actually don’t use Facebook. I don’t have any moral objection to it. I just don’t have the time to do it. [laughs] There are other things of this sort that can end up soaking up a substantial amount of time.
I think that one of the reasons — and this is, of course, quite speculative — I think that one of the reasons people are most worried about Facebook is they don’t really understand the limits of what can be done at Facebook. Whereas at Google, I think we’re pretty clear that we’re showing you ads. We’re showing you ads that are targeted to one thing or another, but that’s how the information’s used.
So, you’ve got this specific application in our case. In Facebook’s case, it’s more amorphous, I think.
COWEN: Wooster, Ohio — I believe you’re from there. Is it economically inefficient? Should its population, over time, be reallocated to larger cities?
VARIAN: It’s funny you mention that because I grew up on a farm — apple orchard — outside of Wooster, Ohio, which is a town of about 20,000 people, and they have a nice college there. The College of Wooster. And it seems to be thriving. So, what happens when you look at these towns in the upper Midwest . . . if they have a hospital, they’ll probably survive. If they don’t have a hospital, they’re in big trouble.
COWEN: In Wooster, Ohio, do you think the value of Facebook and Google, relative to per capita income, is higher or lower than in, say, midtown Manhattan?
VARIAN: We have actually done a little research into this question but only on one aspect, namely, looking at online shopping. And I will tell you, if you live on a farm in the Midwest, you love online shopping. If you’re living in Manhattan, you’ve got a lot of opportunities to go shopping in the physical world. Those rural residents really like the internet for just that reason. The shopping, access to content, all sorts of things.
COWEN: Given how much the world has changed, do our intuitions about antitrust, in any way, need to change? The laws themselves haven’t changed, but obviously, court decisions evolve. But compared to 20 or 30 years ago, what more do we need to bring to that discussion?
VARIAN: I think the existing legal infrastructure and antitrust in the US is pretty good. One of the things that’s very useful here is that the DOJ and the FTC put out guidelines, like the merger guidelines, the intellectual property guidelines, and so on. If you look at the situation in Europe, it’s a less mature system. They don’t provide you with those sorts of guidelines, so you’re flying in the dark to a larger degree.
COWEN: Why has copyright enforcement run so smoothly? And this reminds me a bit of the spam problem. So, the first Napster comes out. You think, “My goodness, no one will ever sell music again.” And I understand the price of streaming is much lower. In part, that’s due to competition from piracy. Yet it’s striking to me how few young people get most of their music from piracy.
VARIAN: Yes.
COWEN: What did we fail to understand early on about Napster?
VARIAN: As you know, there are two very nice recent books on this. One is Joel Waldfogel, who wrote a book called The Digital Renaissance, which looks into exactly the question you were describing. The other book, by Michael D. Smith, is Streaming, Sharing, and Stealing.
I think what they found was that having a well-organized, high-quality, readily available complete access to music was much more attractive than cobbling something together using some sort of back doors and other things of that sort. It’s very clear that people are willing to pay $10 a month to be able to get a very smooth access to the content they want to listen to.
COWEN: Will local news find a business model in the same way that is other than being owned by very wealthy people?
VARIAN: Yeah, this is the thing that surprises me — that nobody seems to be able to crack that nut yet. There’s clearly a demand for local news in terms of just what happened at the high school basketball game or what special offers the supermarket has today, et cetera. People are experimenting with all sorts of different services, and my guess is that something will catch on. It’s a little surprising it’s taken so long for that to happen.
COWEN: What’s your view of Robert Nozick’s entitlement theory of justice?
VARIAN: Well, I knew Robert Nozick.
COWEN: As did I.
VARIAN: He was a friend of mine, and he was, I think, a fantastic guy to talk with and debate with. I would say it was somewhat speculative. I don’t think it would necessarily work out as well as Nozick argued it would work out. There would still be a lot of problems with his theory, in my view.
COWEN: If you could make one reform to the economics profession, what would it be?
VARIAN: I would like to see shorter review periods for journals. I’d like to see things get into print more quickly. I think the whole journal enterprise now has been really bogged down, particularly compared to what goes on in other fields, like computer science.
COWEN: Does tenure still make economic sense? It doesn’t seem to on paper, right?
VARIAN: Yeah. I think it’s going to get more and more difficult because there’s now this big push for enlarging educational access. What does that mean? It means you have to have more capacity in the universities to handle a greater number of students.
So you’re going to see a situation where there’ll be a variety of roles, not just professor tenured and professor nontenured. There might be people that are doing various kinds of intermediary activities as well. Say, tutors, is an example.
COWEN: Last question. Let’s say a very smart 18-year-old comes up to you and says they want to have a career in both economics and tech, but they’re not exactly sure what they want to be. And they want advice from you. Other than the obvious, “Work hard. Go to a good school” — or whatever — what would you tell that person? They say, “I want to be the next Hal Varian.” In some manner.
VARIAN: I would say get the basic skills down. Maybe that’s obvious. Get the coding down. Get the design down. Get the basic economics down.
But then you really want to exercise your creativity. That is, don’t just take the first way of looking at something or the conventional way of looking at something, but try to step back and see what the bigger picture is. Now, most of the time that’ll be a big flop. But every now and then, you’ll hit something that’s new and exciting and novel, and, of course, that’s a great feeling to be able to do that.
COWEN: Hal Varian, thank you very much.
VARIAN: Thank you.