Subscribe and Listen on: Apple Podcasts | Spotify | Stitcher | Google Podcasts
You can find all our Economics and Beyond audio podcast episodes here: https://www.ineteconomics.org/perspectives/podcasts
Transcript
Rob Johnson:
I’m here today with Dina Srinivasan. She’s a fellow with the Thurman Arnold Project at Yale University, trained as a lawyer, she’s worked in technology and developed advertising related technologies, she has been an advisor in recent years for people to understand the kinds of structure that the so called large internet monopolies represent and what it portends for society. Dina, thanks for joining me here today.
Dina Srinivasan:
Thank you for having me, Rob.
Rob Johnson:
I think there’s enormous amount of curiosity and concern I think anytime that technology and the whole way of life is transformed, there’s a transient period of anxiety but particularly when the concentration of the entities in this information world becomes so fierce and so narrow, and for that matter, so different. I know that you have done work, INET has covered, we’ve written papers for our blog related to Facebook and related to Google and you’ve published in scholarly law journals one at Berkeley and one at Stanford, why don’t we discuss both over time over the course of this conversation, but at the outset, let’s talk about what do you see about Facebook and how does it work and what cause for concern or perhaps government intervention in the world of anti trust would, how would I say, address those concerns?
Dina Srinivasan:
Sure. I sort of think of Facebook as sort of the 21st century communications utility. Folks don’t really use landlines anymore, they talk on even mobile phones less, and these social networks have really come up to take over the way that people communicate with each other and share things with friends and family and colleagues. Before I really even started to think about anti trust in a very structured way when it relates to Facebook, one of the things that I thought a lot about was how is it that the people and consumers and citizens are okay with some of the terms that Facebook extracts from them?
Dina Srinivasan:
Now we all think Facebook is free, we sign up for a Facebook account and that’s sort of been static for ages. But actually in 2014, Facebook flipped the switch on one particular thing, one particular term, which I found totally fascinating. It used to be that when you signed up for a Facebook account Facebook can use information that it gleans from your use of Facebook in order to target you with ads, or to just use that data in sort of the ways that it wants to use it. But what happened in June of 2014 was it sort of, consumers that are now using Facebook are agreeing to have Facebook survey their activity across the internet, across thousands and thousands and thousands of sites.
Dina Srinivasan:
So if I wake up in the morning and read an article on the New York times, Facebook can make a record of what it is that I read, what was the title of the article, when did I read it, and that felt like incredibly intrusive to me from a privacy perspective. I became very curious as to why is this a bargain that people actually like? Is it a bargain that people like and how did it come to be that this was the case? So I’m not sure if that answers your question, but that was sort of the origin of my interest in Facebook.
Rob Johnson:
I just want to kind of roll back in time, so the individual signs up for Facebook and it’s, I say, it’s a delightful thing to be able to keep in touch with your high school or college friends and your family that’s now dispersed around the country and so forth, so you feel like you’re getting, how would I say, the ability to shape private networks. What is it that you’re also giving up that you’re not being told you’re giving up?
Dina Srinivasan:
So you’re giving up privacy, right? You’re giving up different levels of privacy. It used to be one level of privacy, Facebook was competing with other firms in the market. When Facebook entered the market in 2004 it promised users we will not track you across the internet, because in part, Facebook was trying to differentiate itself from the very privacy intrusive Myspace, Myspace was a disaster when it came to privacy. And as time moved on, Facebook tried to change that bar again with users, but you have this really long history of consumers pushing back and revolting and saying, “No, we don’t want you to track us across the internet.” And then Facebook saying, “Okay, golly, shucks, we won’t do that.” And that sort of narrative went on for about 10 years until other social networks folded, went out of business, exited the market, and then Facebook could finally get users to consent to that. So to me this looked like a very classic monopoly rent problem, it just wasn’t coming in the form of dollars and cents.
Rob Johnson:
Mm-hmm (affirmative). And in essence, I guess Facebook, first of all, doesn’t or did not disclose what they were doing with the data, and secondly it seems like for the data to be valuable it needs to be what you might call large scale aggregations, so that they’re not, how do you say, telling secrets on 25 people, they’re creating clusters of big data that allow pattern recognition or political parties to know which people they should reach out to for fundraising or not, all kinds of different things that might be what you might call parsed based on the data or based on the things that are being monitored. Is that the right way of seeing it?
Dina Srinivasan:
Yeah, in part, data is valuable in as much as it’s aggregated and crunched in aggregate ways like that. But one of the important things to understand and appreciate with Facebook is that it can correlate sort of the stuff that you’re doing with your real identity, because you have to sign up for a Facebook account using your real name. So those data correlations are actually made at sort of a Dina Srinivasan level, right? They’re made at sort off the SSM level practically.
Rob Johnson:
Mm-hmm (affirmative). Now have you seen, I would presume you have, the film Tristan Harris and others made, The Social Dilemma?
Dina Srinivasan:
I actually have not watched that in full yet.
Rob Johnson:
Yeah, Tristan has been involved in a number of INET events based in the presidial in our California office over time, lovely man, very earnest, very interesting. He and Jaron Lanier and others I thought shed a lot of light on these approaching challenges in some of the forums that we’ve been fortunate enough to organize under Pia Malaney. The sense that I got in that film is that there are concerns related from all kinds of internet use, related to brain development and things in children, and the second was the way in which the monitoring of your data and your responses and what you chose to read, allow them to essentially characterize people and then sort out the stimulus so that, how would they say, you were getting positive reinforcement for what you wanted to believe, not so much challenged or contradicted. And this would then boost enthusiasm, spreading of the network, and ultimately increase advertising dollars.
Rob Johnson:
But as I think, I saw this film quite a while back, but there was some notion that by creating this filter bubble as Eli Pariser wrote about, you created this bifurcation that “ran the danger of creating or fomenting a civil war in the United States.” By positively reinforcing people to energize their evangelism for joining the network and their participation in the network, we were fomenting a divide that could be very dangerous politically.
Dina Srinivasan:
Yeah, definitely. I mean I think sort of the elephant in the room with Facebook and other large companies that rely on digital advertising for their revenue base is that they need you to stay on the platform for longer and longer periods of time because they make more money the more time you that you get into platforms, and then they made more money based on the more that they target content and target ads to you, and that incentivizes them to basically collect more and more data.
Rob Johnson:
So at this juncture, there are lawsuits that have been filed vis-a-vis Facebook in a number of places and what do people envision is the time table? Are these lawsuits … I get the impression, I’ve talked a little bit to Roger McNamee and some others that it’s almost incomprehensible whereas there are a number of people whether in law or in economics trained in dealing with traditional monopolies, the novelty of the structure and the patterns and where the violations of social good occur, the kind of things that you have shed light on, how do I say, they’re not yet digested to the place where we understand in a court of law how to pursue right and wrong. Do you sense that the world is ready to hear these cases? I feel a sense of urgency, I feel an instinct, but to make wise judgements and decisions, is that in the offing?
Dina Srinivasan:
I think we’re totally ready to hear those cases. I think the cases actually themselves do not present such novel questions of fact or law for courts and for judges. I think really the incredible thing, and the really hard thing which is now behind us, is actually the filing of the cases, right? Because you have to have a lot of … it’s a lot. You’re filing a case that is novel and you’re filing it really to in large part restore competition in order to restore better sort of products to consumers. You’re filing these cases to restore for consumers communications channels that offer them greater levels of privacy. And so I think it’s just an incredible thing that you have governments that are using tax dollars to wage these fights on behalf of the people for the benefit of their privacy, and I think that’s where the really heavy lift was.
Dina Srinivasan:
You can argue to a court, look this company is a monopoly because it’s big and it has a large share, and that’s sort of a standard argument that you can make for Facebook as well, and then you move onto the argument, well this company is a bad monopoly and it’s harming consumers because it charges monopoly rents, it charges much higher prices, you can make the same argument when it comes to Facebook, instead of saying it charges high prices, you say it’s reduced the quality of it’s goods. So there’s sort of very standard arguments that are being made, but they’re just being made through slightly new lenses.
Rob Johnson:
Okay. What took place in the question of you want to call mergers and acquisitions, were there any stipulations made when WhatsApp was acquired by Facebook or when they coupled with Instagram? Do people see that as, how do I say, a type of consolidation that reinforces the monopoly power? Or is that a different thing in this instance?
Dina Srinivasan:
Sure. So a lot of folks see the acquisitions that Facebook made of WhatsApp and isn’t as problematic from a monopoly perspective, and one of the things that I sort of grapple with here or that a lot of folks grapple with is okay, suppose we break up Facebook from WhatsApp tomorrow, what is the world going to look like in 30 days? In 30 days or in 60 days or even 90 days, the billions of people who have Facebook accounts will stay on Facebook, and the billions of people that have WhatsApp accounts will stay on WhatsApp, and they will have two accounts. They’ll have a Facebook account and they’ll have a WhatsApp account, because the two products really are not interchangeable. One is a messaging app, one is a social network app, and you use both, you don’t use one or the other.
Dina Srinivasan:
So that is to say that even if you go after these mergers, consumers are still going to use both products after you break them up, and you’re not really introducing choice or competition in the market. But the very good argument here on the flip side of this is that yeah, if you had other companies on the market, if you had a separate Instagram, you had a separate WhatsApp, and they each had a user base of hundreds of millions or billions of users, those companies in the near future could decide to retool and compete with Facebook directly. So it’s that possibility of competition that could help the market, which is why the cases are going after those mergers.
Rob Johnson:
Mm-hmm (affirmative). Okay. Now recently you’ve written a paper, as I mentioned at the outset, that INET helped support related to the concerns about Google and particularly as it relates to the advertising model. When I read your article on the INET website and looked through the paper, I saw that you drew analogies to the electronic trading world in financial markets. Let’s explore with our listeners, what do you see, what is the structure and what is going on with Google that you chose to shed light on?
Dina Srinivasan:
So this story really begins in 2014. I was reading Michael Lewis’ excellent Flash Boys book, and executives in the digital advertising industry were passing that book around and sharing it with each other and then telling stories of, “Ha ha this happens in our market too.” And amongst the business crowd, it wasn’t such a spectacular thing that you had this thing in Flash Boys that was happening in the ad industry, but for me with my sort of, I don’t know, I was trained in law, I was trained in competition issues and policy issues, and so I just thought it was fascinating. I was like wait a minute, the whole world is going crazy about this narrative in Flash Boys, if it’s happening in advertising and nobody knows and nobody’s saying anything, that’s crazy. That’s really interesting to have that just regulatory vacuum in the ads industry.
Dina Srinivasan:
So last year when I was sort of breaking down what to, actually this was in the summer of 2019, what research to write about next I decided to take on a comparison between the ad markets and financial markets. And at a really high level we’re talking about the online advertising market, so we’re talking about those online square box ads that you see when you log onto a news site for example, you go to the New York Times, you pull up the site and there’s a little box ad in the top right hand corner at the top of the page. Those are called banner ads, and those banner ads today, the vast number of banner ads, maybe about 86% of the entire market in the US is traded in real time on centralized electronic trading venues that the industry calls ad exchanges. The industry really migrated to electronically trading these things on centralized exchanges in around 2004 which is the same year really that the New York Stock Exchange migrated to electronic trading as well.
Dina Srinivasan:
So you have billions and billions and billions of ads changing hands on these centralized exchanges, and then the structure of the market from a broader perspective also looks very similar to financial markets in that buyers and sellers also have to go through a broker or middle man to trade on these exchanges. So if you’re selling ad space, if you’re the New York Times or another website or even an app, you have to go through an intermediary in order to sell your ad space on exchanges. Now these intermediaries in the ads market have funky names like ad servers, but they’re basically the same thing. Your ad server decides all right I’m going to route your ad space to this exchange, I’m not going to route it to another exchange, I’m going to route them sequentially to different exchanges, they made decisions on your behalf about how to route space to [inaudible 00:20:56]. And then on the buy side, you have advertisers who also have to go through a middle man in order to buy ads on exchanges.
Dina Srinivasan:
So from a macro perspective, the structure of the market looks similar to the structure of financial markets if you want to buy a share to Tesla you can’t go knock on the door of the New York Stock Exchange, you have to go through a broker.
Rob Johnson:
So you have a selling broker and you have a buying broker and then at the center is an exchange, and yet I recall from reading your work that sometimes all three were owned by Google.
Dina Srinivasan:
Exactly. Exactly. And that’s one of the fascinating things about this market, and really would be a fascinating thing about any exchange market, right? It’s not just stocks and ads that trade on exchanges, but you have other sectors of the economy as well that have migrated to electronic trading in this fashion. Which to your point, yes, almost any ad that you see you have a very, very high likelihood that Google represented the seller and routing it to Google’s exchange, it cleared on Google’s exchange and Google was sort of on the buy side on behalf of the advertiser as well at the same time. It’s wearing all three hats at the same time.
Rob Johnson:
Mm-hmm (affirmative). And the other thing I remember when I read your paper, unlike financial markets, the price of advertising for the buyer and the commissions for the exchange and the brokers is not transparent, it’s not disclosed, is that correct?
Dina Srinivasan:
So, it gets a little bit complicated. But oftentimes what will happen, I’ll just give you one example here, at the big picture level there’s a huge problem with transparency, and as we know with our experience regulating financial markets, when you don’t have transparency you combine lack of transparency with sort of light conflicts of interest in a complex electronically traded market and you have just disaster waiting, right? So you have the same thing in the ads market, and one example is for example, if you are a website, you’re selling your ad inventory and you’re using Google on the sell side as sort of your broker, and by the way, Google has over 90% share on the sell side, so everybody uses Google on the sell side, and then let’s say Google preferentially is routing that space to Google’s exchange, so now it’s clearing it on Google’s exchange. At the end of the day, Google’s exchange gives data dumps back to the websites and it says, “Well here’s what your ads cleared for. This was the clearing price for all your goods.”
Dina Srinivasan:
But when it does that, it’s not necessarily disclosing that, oh this car dealership paid X price for your ad and that was the clearing price on the exchange. The disclosure is at the buy side sort of broker level. So what did Google pay on the buy side when it was wearing it’s buy side hat, what did Google pay for that ad on Google’s exchange? So you know what the ad cleared for, you have transparency to the clearing price, but you don’t have transparency into for example Google’s margin when it’s acting on the buy side. I know that was very complicated.
Rob Johnson:
So I know what I paid in gross, but I don’t know how it gets cut up between all the different intermediaries between there and the end, that’s the interesting-
Dina Srinivasan:
Exactly.
Rob Johnson:
That’s interesting. Now I remember also from reading your paper that over time there’s been, where there used to be many exchanges, I remember Yahoo had one and there were others, but that’s become, how do I say, I guess what you might say the volume dried up in some of them and they went out of business or left the business, and the Google exchange has become what I’ll call, again, the natural monopolists. Had the prices and margins which I guess are not disclosed, so maybe I’m asking a foolish question, but have they come down? Is there a sense in which by that natural monopoly getting all that kind of volume with economies of scale, is there any notion that the price of the matchmaking service or the gross price for advertising has come down? Or is that just too difficult to disaggregate or discern?
Dina Srinivasan:
No, I think actually the prices for I think [inaudible 00:26:27] have gone up, and Google’s exchange charges more than other exchanges in the space, and those margins have been incredibly very interestingly resilient to sort of any price pressure. Going back to that story, that story is really interesting. How is it, you had a market where it started developing in 2004, and you had Yahoo was operating the number one exchange in the space, I think Microsoft was operating number two, you had lots of exchange competitors, and Google decides to launch an exchange in 2009, like five years after the development of a market when you already have a bunch of players. And the question is how did Google sort of enter a competitive exchange market in 2009 and very expediently just push out the competition, right? And the fascinating thing for me at least with this Google research is you look at sort of all of Google’s conduct in the advertising market, and things look random, it’s doing this, and it’s doing this, that’s the way it looked to me. Then it’s doing this and I didn’t really know how to make sense of it all.
Dina Srinivasan:
But after studying financial markets you realize, oh every single thing that Google was doing is something we prohibit in financial markets. It’s not random at all. So the way that Google’s exchange sort of entered a competitive market tells one of these stories. So Google bought Double Click, which was the number one sell site broker, and then it adopted this rule which said that when we’re routing websites face to an exchange, we’re going to preferentially route it to our own exchange. Not necessarily if the price is higher or better for the seller, but there are just these preferences of routing order flow to Google’s exchange. So because it controls the sell side, it was able to do that and starve other exchanges of liquidity and they just went out of business.
Rob Johnson:
How would I say, new technology world. Another company that we see in creasing in skill very rapidly is Amazon. It probably is a different type of model, but do you have peers or colleagues or people who are exploring the anti trust implications of the growth of Amazon.com?
Dina Srinivasan:
Sure. I think Nina [inaudible 00:29:27] here has been really revolutionary in that regard.
Rob Johnson:
And are there similar kinds of stories or structures that are being revealed there? I get the impression that Amazon has many dimensions. They have what you might call the retail at your door, or books and music and clothing and what have you, but I gather they’re also cloud systems and other things that they’re in the market for, or market of providing. I was just curious as to whether there seems to be as much angst related to them as we’re experiencing vis-a-vis Facebook and Google?
Dina Srinivasan:
I think the angst is definitely there. I think that the angst and the economic puzzles presented by each of these companies is very different, the market structures are very different, the things that they’re selling are very different, the relationship with consumers are very different. So you have a different set of issues. However, I would say that there’s some issues that tend to overlap, and one issue that tends to overlap that I think, I spent a lot of time thinking about and I think is just really interesting is around the use of data. So when it comes to Google, when Google is operating the exchange or it’s operating on the sell side, when it’s basically wearing different hats, it doesn’t keep customers data, so the actual traders in the market, it doesn’t keep their data cordoned off behind a wall for this use and then this data for this use, it tends to merge data behind the scenes and it can trade on that data.
Dina Srinivasan:
Now, because the ad market, the structure of the ad market is so similar to the structure of financial markets, you can then look at that puzzle and understand the economic inefficiency behind it because you say, oh that’s just like insider trading, right? That’s what’s actually going on here, that’s the economic problem. So when it comes to Amazon, one of the things that came up, I believe it was last year, was around Amazon using the data of third parties to inform what products it should create, manufacture, sell, and how to sell it on it’s own platform, right? So whereas you have completely different businesses and different business models, that’s a conflict of interest problem and it’s a conflict of interest when they’re wearing different hats, they’re using data of their competitors just to sell their own products, and it feels like it’s looking very similar to sort of this insider trading problem.
Rob Johnson:
I can imagine the pressure, let’s say if you were a book publisher, to have to give Amazon the best price could be quite, how do I say, quite powerful, quite daunting in the sense that if you didn’t do that they would play such a big role in the market that, I won’t say dooming, but you’re impairing the quality of sales. When I worked in music it was often the case that people would draw inference from sales. This was the era of CDs, but if a CD looked like it was selling well in places analog to Barnes and Noble, Tower Records in those days, would all be incented to buy and put in the window these things that they knew were the “hot products” and it would be very hard now I think for a book to be considered a hot product if it was aggressive in it’s pricing vis-a-vis Amazon relative to other outlets. So there’s a lot of different dimensions.
Dina Srinivasan:
Dimensions.
Rob Johnson:
Let me ask you a question, I know you’ve worked closely with our research director Tom Ferguson, and he’s a profound contributor to the question of the role of money in politicts. To my mind, one of the most striking things about this modern era, we’ll call it from the mid 1970s to the present, is the almost absence of anti trust investigation and enforcement. At this juncture, with highly concentrated wealth, large corporations treated as persons, are we in a place where which you might call the dis incentive for an administration to enforce and investigate, bring to trial and enforce anti trust has become, how do I say, that disincentive has become very strong for fear of losing donors or in the case of a Google or a Facebook, having very powerful, high volume marketing instruments.
Rob Johnson:
I have this sense from Tom that in many, many respects, government, whether it’s legislation, appointments, enforcement of existing rules, are let’s just say refracted by the maintaining the focus on building the war chest for maintaining one’s position in elected office whether it be the president or speaker of the house or the head of the justice committee in the Senate or House, is there a sense, is my intuition from reading Tom’s work that the pursuit of anti trust might be somewhat less vigorous because of the awareness of the role of money in politicts?
Dina Srinivasan:
Well I would hope that the pendulum sort of swung far enough in that direction, that’s one of the reasons that maybe, I hope it is sort of bouncing back and at the end of the day, I don’t know, I’m very hopeful that these anti trust issues and questions that have been raised against Google and Facebook and Amazon have drawn such bipartisan support, so I guess I would say that that’s good news, and let’s hope that type of good news continues. Because I think they present really [inaudible 00:37:22] historically important questions that are much bigger and much more important than the issues you raised.
Rob Johnson:
I’m curious, at this juncture going forward, can you share with us what kind of next projects you envision? What kind of things are on your radar screen right now with your interest in anti trust beyond Google and Facebook, do you have other innovative platforms or environments that you would like to investigate?
Dina Srinivasan:
Well I’m itching and yearning to figure out what’s next. One of the things that I do when I finish a project, and just sort of figure out what to do next, is I try to just clear my calendar as much as possible and take a step back and read, read all the papers to sort of have one leg in current events and have different perspectives of current events, but then I just read a stack of economics books that I have on my agenda to read, and then I try to start reading a little bit of fiction again. And then that sort of [inaudible 00:38:51] I think of something else, but I’m still in that stage. So I don’t know. But I would imagine that it’s still going to fit in the bucket of research and sort of writing and then trying to work on these issues in a real world way to effectuate change.
Rob Johnson:
I’m just, knowing you’re working with the Thurmond Arnold Project at Yale, have you read his book, The Folklore of American Capitalism?
Dina Srinivasan:
I have not.
Rob Johnson:
Or I guess it’s called the Folklore of Capitalism. That’s quite a good read. I remember reading that when I was a graduate student. It really got under the skin of some of these questions about concentration of economic power. I thought that was, I don’t even know if it’s still in print, I think it would be quite interesting to, how do you say, in light of the pathway that you’re on to visit his thinking and some of the earlier, what I’ll call progressive approaches to anti trust enforcement. Though as you’ve pointed out nicely in your research, these are different structures. It’s not the traditional model.
Rob Johnson:
I think that, how would I say, the final thing that I really want to emphasize for our listeners is the extraordinary contribution of people like yourself. We all, as HL Menkin wrote about in his famous 1920s article, The Dismal Science, he said that all social analysts and economists are mindful or intuitively mindful perhaps of the ramifications of challenging power, and many people, I know Gillian Tett who’s on the board of INET who works with the Financial Times, once told me that you should study what’s not said because the map of the silences tells you where power is. And what I’m coming to is you have chosen in your own career as a mother of four children, as a citizen, to illuminate and to challenge what these new structures are doing or could do, both positive and negative, for the quality of our lives. And there’s some very concentrated and powerful interests that HL Menkin would point to that are on the other side of that.
Rob Johnson:
So I guess what I’m trying to say in a long winded way is thank you for the courage that you exhibit for your perseverance, for the intensity of your intellect in looking these challenges right in the eye. You are an example, my young scholars should pay very close attention as they’re developing their careers, because you are an example of someone who is providing for the public good. And I really want to applaud that.
Dina Srinivasan:
Well thank you so much, Rob. It was just so generous.
Rob Johnson:
How would I say, I think we need more of that right now. There are a whole lot of things that are changing, there are a whole lot of things being unmasked, and in addition to the illumination that you create, the example that you create is a call to action for the rest of us to pick up and embrace our work in social science in a similar manner. Any last thoughts that you’d like to share with the audience before we sign off?
Dina Srinivasan:
It’s been a pleasure talking to you, thank you so much for having me, and thank you for doing all of these podcasts. I really enjoy listening to them.
Rob Johnson:
It’s my pleasure and I’m very grateful that you could make the time. I hope at some point in the not too distant future as these issues continue to unfold and you continue illuminating them that we might join again and do another episode of Economics and Beyond.
Dina Srinivasan:
That sounds like a plan.
Rob Johnson:
Excellent. Bye bye for now.
Dina Srinivasan:
Bye bye.
