My guest today, Susan Athey, is not only a superstar academic economist who’s on the short list of likely future Nobel Prize winners, but also someone who’s having a huge real-world impact.
ATHEY: I was familiarizing myself with not just one problem of the future, but, like, 10 problems of the future. And I was one of the first economists in the world to see it.
* * *
Welcome to People I (Mostly) Admire, with Steve Levitt.
* * *
Susan Athey is a professor at the Stanford University Graduate School of Business, and in 2007, was the first woman to win the John Bates Clark medal given to the most promising American economist under the age of 40. She was chief economist at Microsoft for six years, played a key role in the U.S. government’s Covid vaccine strategy, and now is chief economist in the antitrust division at the U.S. Department of Justice. And separately part of a nearly $1 billion plan to accelerate a market for pricing carbon.
LEVITT: You’ve made, what I would say risky choices over and over. And the first one is that you spent a number of years working closely with Microsoft. And nowadays it’s pretty common for top academic economists to work with private companies, but when you did that back in the mid-2000s, it was surprising and unusual. What were you thinking at the time?
ATHEY: So, it was actually a moment when I was looking to pivot. And in fact, I had made the decision that I wanted to start thinking about health, and I was thinking hard about doing a shift to something that would have a big impact on the world, and that I could use all the skills I developed to date. But then, along came this bizarre, completely out of the blue opportunity, where I get this phone call and say, “Steve Bomber would like you to come talk to him.”
LEVITT: And Steve Ballmer was the C.E.O. of Microsoft at the time.
ATHEY: Yeah. But I never met him. I didn’t even know what the meeting was going to be about.
LEVITT: How did he know who you were?
ATHEY: He read an article about me in the Harvard Alumni Magazine. And so, he was at the time struggling with competing with Google. He read about me, and he saw that I was an expert in auctions. Google was running auctions for advertising. He realized he wanted some new and fresh thinking in that area. So, he brought me in and basically just said, “Hey, we have all these problems. We need to think about how to compete with Google. We’re looking for some new thinking. What do you have to say?” It was really open-ended. Incredibly open-ended. And I had no experience whatsoever with consulting for companies.
LEVITT: But they hired you, and you had a long and I suspect very fruitful relationship with them.
ATHEY: Well, there was a lot of negative feedback from my academic colleagues because at the time the idea was if you left full-time research for anything, you might not come back. And the idea was that if I did this, I was leaving research forever. But I looked at the opportunities that I saw in the questions. And I realized that my brain was getting challenged as hard or harder than it had been doing pure research. And in this particular environment, there was a couple of really unique things going on. The search engines were really different things. And in the mid-2000s, most folks outside of a few industry insiders did not really understand what this was all about. I looked at it and I saw there was artificial intelligence. Underlying that were machine learning algorithms, which I saw had the potential to revolutionize the way that we did social science.
There was an A/B testing platform where thousands of randomized controlled trials were being run on a regular basis, where, in academics, we were occasionally running these small-scale experiments. These were large-scale experiments on real humans. And then the thing itself, the search engine, was getting in between people and their information and that had the potential to affect the news we read and what companies win and lose in e-commerce, and all sorts of other things. The information getting collected had privacy and security implications. So, just within a couple of months, it was completely overwhelming — the number of incredibly important issues out there. And then people were coming to me with questions and I realized there was very little academic literature to build on because most academics didn’t even understand what these problems were. I was familiarizing myself with not just one problem of the future, but like 10 problems of the future. So, this was clearly a multi-decade research agenda. And I was one of the first economists in the world to see it.
LEVITT: You it’s funny because I think the profession as a whole has come to understand that these partnerships with business are compliments to doing great academic research, not substitutes. But it took pioneers like you to take those first steps, to show the rest of us just how rich this universe was that we hadn’t tapped into.
ATHEY: And, of course, there were some really important challenges in doing this. I was there pretty intensely for about five years. But one reason it lasted that long was that Microsoft had this Microsoft research division. So, I could have a foot in business, but also there were researchers who were actively engaged in publishing. So, there was the ability to be a part of research, even with the work I was doing at Microsoft. But ultimately, the challenge you face is that as an academic, you have to get up close with these companies to really understand them. But at the same time, you want to be able to stay objective. That’s a pretty challenging path to walk.
LEVITT: You mentioned machine learning and artificial intelligence in the context of Microsoft. Is that where you first got introduced and interested in these ideas, which were really quite foreign to economists at the time?
ATHEY: Absolutely. I came to Microsoft in 2007 with expertise in auctions using data the way that economists use data, which is to ask questions about cause and effect, and to estimate what we call counterfactuals. So, that was an expertise that I had. There were no economists, basically, at Microsoft. But there were hundreds or thousands of people, working on machine learning for the search engine. And they were using data, but using it in a very different way. The off-the-shelf machine learning that was used at the time were predictive models. So, you have data and you’re trying to say, classify, is this picture a cat or is this picture a dog? Or you’re trying to classify a query and say, “Does this query have something to do with movies or weather or is it a good match for Wikipedia entries?” The whole system was built up with a bunch of classification algorithms.
And so, the machine learning experts that were really good at that generally had not thought much about using data for changing the world. Their tools were not necessarily well set up for saying, “What would happen if I changed the auction rules?” Or, “What would happen if I introduced a new advertising product?” Those types of counterfactuals weren’t something they were used to thinking of. And so, it made me realize that there was actually a huge hole in the literature between the algorithmic innovations, the major algorithmic innovations that the machine learning people have been making to use data at scale for things like classification with the tools that we’ve been using in economics to try to use observed data to quantitatively assess what would happen if we changed the environment in some way.
LEVITT: The most recent endeavor that I’ve seen your name attached to — and that’s a billion-dollar philanthropic project, providing advanced commitments for carbon capture. But before we go into the specifics of that project, could I just get you to explain what advanced commitments are and why they might be a good idea?
ATHEY: We often have innovation that’s going to benefit society. And people have to make expensive investments upfront. And they aren’t really sure whether there’s a consumer on the other end for the product, because the product is new. And complicating things, sometimes the product is going to be purchased by governments. Anytime you’re selling something to a small number of buyers, you have to worry about exactly how that negotiation is going to go, but governments, of course, are special. So, when you’re in this kind of uncertain environment, you might end up with a lot less investment than would be socially beneficial. And various types of advanced commitments are a way that we can help reduce uncertainty and get more valuable investment.
LEVITT: So, an advance commitment means that your buyer, maybe the government, maybe some other private company, before you’ve actually made the product, says, “If you make such and such a product, then we will pay you this many dollars for your product.” And that’s to get around what economists call holdup, where maybe you make this huge investment and then it turns out the government says, “I’m going to pay you less than we would have agreed to, if we had bargained before you made that big investment.”
ATHEY: That’s right. You’ve seen these come into play with vaccines, and there can be a lot of pressure on governments to get a good price for vaccines. And the drug companies are not always the most sympathetic firms, even if they’ve made something that’s really beneficial for society. And so, the companies might, reasonably, worry that after they’ve built a factory and they’re ready to produce vaccines that the government — they have the authority to just say, “Well, you’re going to make them, and this is how much we’re going to pay.”
LEVITT: If the pharma company had known ahead of time that the government would only pay, say, $7 a dose, then they never would have invested the billion dollars in doing the research to create the drug. But if they knew the price would be $15 a dose, then they would have done it. And this is a way of the company saying, “Hey, I don’t really trust the government, but if they’ll agree up front to pay me $15, then I’ll go ahead and take the risk and do the research.”
ATHEY: And it’s natural for consumers and citizens, after you have this great product, to want it for the lowest possible price, but a lot of these facilities that were ready to produce vaccines were making something else that was profitable. And so, in order to get as much capacity as you’d like, you need to make sure that the companies are willing to move heaven and earth to change over to something else that might be less profitable. And so, understanding the financial side of that is really important for understanding what it’s going to take to really induce the companies to make the decisions you want them to make. If they’re anticipating that they might end up getting paid a little later, they might just say, “Well, sorry, that facility isn’t available.” And, “Oh, it’s too hard.”
LEVITT: Now the first big advance commitment that I’m aware of was for vaccines, but it wasn’t the one people know about — about the Covid vaccines. It was well before that, for pneumococcal vaccine? And my sense is that turned into a huge success. Am I right about that?
ATHEY: That’s right. So, way back in the late ‘90s, Michael Kramer was a junior professor at M.I.T., and I actually happened to be his neighbor on the hallway. He said, “I’m going to cure malaria.” And I said, “Huh? You’re an assistant professor of economics and so am I. How are you going to have an influence on curing malaria?” But then he laid out how different types of commitments like this could potentially increase innovation and contribute to a cure for malaria. Now he went out and talked to a lot of companies and looked at where the research was. And along the way, discovered that while malaria was still pretty challenging, there was this pneumococcal vaccine that had proceeded far enough in development that you could be fairly confident that you would be able to get this vaccine to be developed for developing countries where there was a huge problem and many deaths.
And so, he, and some colleagues, Chris Snyder and John Levin and Rachel Glennerster, really over a series of years, put together a very concrete proposal for how to do an advance market commitment for this pneumococcal vaccine. And basically, what they needed to do was subsidize the doses so that the pharmaceutical companies could be confident that if they produced a vaccine that developing countries wanted to buy, there would be a market. This was, though, an advanced market commitment. It wasn’t aimed at a specific company necessarily, and this had some nice properties because if one company came along and had a vaccine that was easier to administer or easier to store or required fewer courses, or had less side effects or anything else, that vaccine would be demanded more, and they would therefore sell more. So, they had an incentive to really create not just a minimal vaccine that met the basic criteria, but one that would attract demand. It gave good incentives for innovation and quality, as well as just in getting something that met the minimal criteria.
LEVITT: And then it worked out. The fantastic vaccine was developed. It was widely implemented, and I can’t think of many cases where academic economists have had as positive an impact as they did in this particular setting. But no one’s ever heard about it, which is interesting.
ATHEY: Well, it’s very well known in the public health community, but less well known broadly. It’s a little bit hard to know counterfactually what would have happened without this advance market commitment, but the best estimates are that it saved about 700,000 lives by getting the vaccine to market at least five years sooner than it would have otherwise. And of course, there’s no guarantee it would have come to market at all.
LEVITT: Now, that is a case where there wasn’t really a viable market for this vaccine because the customers couldn’t pay very much. And the advanced market commitments served a specific purpose. Maybe I’m wrong, but my sense is the other place where we’ve seen this work, and work really well, was with Covid. But the use case there seemed very different in that I have the feeling that the big pharma companies were actively engaged in trying to develop a Covid vaccine, but the real purpose of the advanced commitments there by governments, like the U.S. government — billions of dollars — was not to make sure that this vaccine was made, but more to get priority, to get in the front of the line to try to get the scarce vaccine given to them first. And maybe that’s not exactly right, but that’s my sense, that those are two very different use cases. Do you agree? Or am I off on that?
ATHEY: It’s a little complicated. So, let me start with how I think it should have worked. And then we can talk about how it actually worked. So, early in the Covid pandemic, Michael Kremer and Rachel Glennerster, Chris Snyder, convened a new team, which I was on that tried to help advise governments and N.G.O.s on how to do the best we could to ensure that we would get the vaccines out as quickly as possible. It seemed like in this case, that the pharmaceutical companies would be more confident that there would be a market for this vaccine. But I do want to take us back — I know it’s hard to imagine and remember now, but in April 2020 — “Could there be two waves? Oh my goodness. No, we can’t possibly fathom two waves.” So, there was substantial uncertainty about how long the pandemic would actually last, whether it would burn itself out due to herd immunity fairly quickly.
And then the best estimates from inside the industry were that it was most likely that it would take a year or more to get the vaccines out. The mRNA was first show for this technology and the production had never been scaled. And just to keep in mind that, typically, the life cycle for vaccine R&D from the beginning to at scale delivery is closer to 10 years. And really no one had ever gotten something to market that quickly before. So, this was unprecedented and there were a lot of unknown unknowns. When we spoke to all the best experts, they felt it was theoretically possible that you could get vaccines in seven or eight months. It was just very unlikely that nothing would go wrong along the way. And if you look back and see all the things that went wrong, what you may not realize is we are really lucky. So, really the chances that they would get this done before the pandemic ended were not a hundred percent.
And then, of course, there’s a lot of demand for the vaccines. But what’s the counterfactual? You build the vaccine, you build the factory, and then at the end, you hold an auction among governments and hope to get a really high price. Well, that’s unlikely that the governments will actually let that happen. Governments ultimately can just confiscate them using Defense Production Act or similar laws. I don’t think the companies really could anticipate exactly what terms they would get on the other side, even though it seemed like something that would be really scarce and everybody would want it. Given the institutional environment, it’s not quite so simple.
LEVITT: And it also probably gave the pharma companies the confidence to just massively scale right at the beginning, and they knew that really early, I think. They knew the mRNA was going to work well, that they just needed to focus all the efforts on as much production as possible. Probably, they wouldn’t have done that absent these commitments.
ATHEY: That’s right. And so, let me then come back to what we recommended, actually very early on, was that we realized that one of the main bottlenecks was going to be capacity. And when you looked at the economic benefits of getting the whole world up and running faster, relative to the costs, we estimated the social benefits could be as much as a $1,000 a dose or more in a developed country, if you were in a situation where your health system was overwhelmed. And so, in that type of environment, the social benefits of getting people vaccinated quickly were really high. And once you’ve negotiated a price, there’s still another conflict of interest between society and the companies. And that is: the companies don’t care as much as the rest of us do about how fast they produce. So, if you think about having double the capacity, that gets the world vaccinated twice as fast. It costs twice as much. But in an environment of scarcity, it doesn’t actually change the revenue for the company.
And we also understood that there was really scarce capacity. Takes a couple of years to get a facility ready to produce vaccines from scratch. And so, you’re going to be repurposing existing facilities, many of which were already in use for other purposes. And so, we recommended that these advanced commitments directly contract on capacity to expand capacity faster. Now, what actually happened was there were these advance purchase commitments. We don’t know all the details of the contracts, but to my best estimate, they didn’t contract as much on capacity as they could have. And then they gave the outside impression of being more about getting earlier in the queue. That’s the way they were framed and discussed publicly. Which then led to a lot of backlash, I would say, where it appeared that rich countries were just jumping to the front of the queue.
If there’s really a fixed pipe, a fixed amount of capacity, then these advanced purchase commitments are just about getting there first. But in practice, I believe that these advanced purchase commitments did have an impact on the amount of capacity and if those commitments did increase the overall capacity, they can actually make everyone in the world better off. Because if you imagine having a skinny pipe that’s supplying water to everyone or a thick pipe, if somebody pays to build a thick pipe and asked to be first in line, it can still be the case that everyone is served faster than they would have been if everyone was sharing a skinny pipe. And so, really the public health positive externality of these commitments is that the rich countries helped finance a thick pipe that could then serve everyone else.
LEVITT: Yeah, absolutely.
ATHEY: But, our recommendation was that pipe should have been even thicker, and it would have been worthwhile to make it even bigger and serve the whole world much more quickly. And unfortunately, we weren’t able to coordinate on a global response that did the very best thing, but that doesn’t mean that the alternative of just not doing any advanced purchase commitments at all, would’ve made people better off.
You’re listening to People I (Mostly) Admire with Steve Levitt and his conversation with Susan Athey. After this short break, they’ll return to talk about Susan’s impact on the carbon removal industry.
* * *
LEVEY: Hey, Steve. So, you’re moving to Germany.
LEVITT: Well, hopefully not permanently — tentatively for a year.
LEVEY: Tell me how this came about.
LEVITT: When I met my wife, Suzanne, she lived in Europe and I lived in Chicago and we made a deal. If she would relocate to the United States for five years while my teenage children were going to high school, when they graduated, I would pay her back by moving to Europe for five years. And I have to say, I thought I would never have to pay that debt back because I thought the U.S. is such an amazing country. How could anyone come here and not love it and want to stay forever? And I misjudged a little because she hasn’t come to love the United States quite as much as I had hoped.
LEVEY: So, a lot of our listeners wrote in and were really excited for you, but it doesn’t sound like you are thinking of it as quite the fun adventure that our audience is.
LEVITT: I’m scared. I mean, I’ve made a big mistake, which is I have not invested nearly enough in learning the language. I was a really diligent student of German for many years, but somehow when Covid hit, I just had the feeling I wouldn’t be going to Germany for a long time and I stopped studying and I never got started again.
LEVEY: But don’t you think living in a country and being immersed and surrounded by the language, you’re going to pick it up a lot more quickly than you would living in Chicago and just using Duolingo to practice.
LEVITT: That’s true. And there is a hidden upside to all of this. One of the problems I have in the United States is that whenever I’m in social settings, I’m almost never anonymous. People know I wrote Freakonomics or I have this podcast and everybody treats me like I’m really special. They give me a lot of deference and they’re nice to me. There’s something wonderful about just being treated like the schmuck that you are. Most people in Germany don’t know what Freakonomics is. And instead of being the person you might want to sit with at the dinner table, I am the worst possible draw. The other thing that’s nice about it is I can go to parties and I’m a real loner. I like to just not talk to anyone, and I can go an entire night and I don’t have to say anything as long as I laugh when other people laugh. People don’t realize I have no idea what’s going on. I love having no obligation to be part of a conversation.
LEVEY: Do you want to know what I’m worried about with you moving to Germany?
LEVITT: I do.
LEVEY: I’m a little worried about losing the amazing audience we have. Since we are going to a biweekly schedule, it makes me a little nervous that we just won’t have the vibrant comments and community that I think we’ve created in the past year and a half of PIMA.
LEVITT: I could see to people who are busy, it would be kind of a relief not to have a whole show every week to have a little more slack time to listen when you can. We’ll find out — it’s an empirical question whether this will lead people to forget about us and lose interest or whether it will actually lead to people listening to a higher share of the episodes than would’ve been true otherwise.
LEVEY: A good thing is that we have a few thousand listeners who live in Germany.
LEVITT: Yeah. And I think about half of them have invited me to dinner while I’m there. It’s been amazing the kind reception that I’m getting from our German listeners, which is great.
LEVEY: I thought a lot of our listeners were going to write in and be disappointed that we were going to a biweekly schedule and they were going to get less episodes, at least, temporarily, but instead everyone was really supportive of your move, and you just got invited to beer gardens all over Germany.
LEVITT: I was going to say if I accepted every invitation we’ve gotten, it would be the most busy social calendar that I had ever had.
LEVEY: Well, thank you, all, who wrote in and shared your well wishes for Steve. If you have a question for us, we can be reached. Our email address is email@example.com. That’s P-I-M-A@freakonomics.com. It’s an acronym for a show. Steve and I will continue to read every email that’s sent, and we look forward to reading yours.
* * *
In the second half of our conversation, I want to talk about Susan’s big new project in carbon markets and get her take on how afraid we should be of big tech in A.I. Also, within the field of economics, there’s a folklore that attributes very superhuman acts to Susan. I’ve always wanted to know if those are true, and I’ll finally have the chance to ask her.
LEVITT: Okay. So, we’ve had two big wins for advanced commitments. So, tell me about Frontier, which you’re hoping will be the third big win.
ATHEY: After we did this work in the vaccine space, our team was out there, and I think it got a lot more publicity and attention and awareness of the importance of advanced commitments of different types. We have this other huge innovation challenge, which is the innovation needed to spare the earth from the more severe consequences that we anticipate might occur. And many scientists estimate that we’ve put so much carbon in the atmosphere that if we don’t remove it we’re going to have pretty catastrophic consequences.
LEVITT: Because the half-life of carbon dioxide in the atmosphere is so long that even the reductions we do now will not be fully felt for a century or more. So, there’s this belief that we need to find technologies to pull the carbon dioxide out of the air.
ATHEY: That’s right. But recently two things have happened. First, some pioneering technologies have started to get traction and at least make people believe that it’s possible that we would be able to remove carbon dioxide at scale. And second, years and years have gone by, we haven’t stopped emitting fast enough. And now, it becomes really a necessity to push into this direction. But we’re in a similar situation to the vaccines in some degree where the firms face pretty substantial uncertainty about who is going to pay for their technology. But in some ways, this is even harder than vaccines, because for carbon removal, the demand doesn’t really exist. It’s not like you use carbon removal. It’s something that is voluntarily going to be purchased either by governments or by companies as part of their climate commitments, but that’s pretty mushy.
LEVITT: It’s not just mushy, but it’s also the problem that, if the U.S. government buys a mega ton of carbon dioxide, the Russian government and the Russian people get exactly the same benefit as the U.S. government and the U.S. people from that purchase. So, we’re in this tragedy-of-the-common situation where even though everyone wants carbon out of the air, they also would rather have somebody else pay for it, given the opportunity.
ATHEY: Exactly. So, there’s these free-rider problems, “let’s let somebody else do it.” And then another way to look at it is, why are companies in this climate business to start with? A lot of it is about their employees — making their employees happy. And it’s about public relations. It’s about corporate and social responsibility, but those are pretty mushy. And those can change. As companies’ financial positions change, the priorities can change.
LEVITT: Okay. So, let’s go to specifics. So, there’s this new thing called Frontier. What is it?
ATHEY: So, Frontier is an organization that has basically organized commitments from a variety of companies to purchase carbon removal in the future. And so, it, came out of Stripe. Now, Stripe is a financial services company. They actually had a really strong team that was interested in climate. And so, they had set up a program over the last couple of years where customers of Stripe could choose to donate a percentage of their revenue to be used for permanent carbon removal. But in talking to the companies that were actually providing the carbon removal services, they learned that these innovators were facing challenges in raising money to scale out their carbon removal capabilities. So, imagine, you’re building a technology. That technology takes carbon out of the atmosphere, and then you want to go to get a loan. And you say, “Hey, I’m going to build this great product. This product is removing carbon from the atmosphere.”
And of course, the people who are going to give you the loan or make the investment, want to say, “Tell me about your customers and tell me about their demand.” And then the poor entrepreneur needs to spend an hour explaining, “I think maybe, Microsoft might buy some or maybe someday Congress will pass a law and then those people will buy it.” And then the financer says, “Wait a second, you’re telling me it might cost $400 or $500 using your technology. So, what makes you think that anybody’s going to buy your expensive way of meeting net-zero commitments when there are much cheaper ways, like planting trees that also helps companies say that they’re net zero. Why do you think anybody’s going to actually purchase your product?” And since this product is so nebulous, providing social goods for the world, trying to reason about who and why someone’s going to buy it is very difficult. And then they never get through the meeting when they’re trying to get financing for their plant.
So, the idea came about that there was some of this problem with vaccines — these advanced market commitments or advanced purchase commitments. So, if you talk to some of the corporate philanthropists about how they’re planning to deploy dollars to meet net zero commitments, they might like to invest in carbon removal, but because they haven’t made any commitments to it, the companies that are trying to create the technology can’t convince the financiers that they actually do have customers.
So, the idea of Frontier was to take those corporate philanthropists who were willing to invest, and get them to go ahead and make the commitment. The argument being: just because you were planning to become a customer of carbon removal in the future, you knowing your plan is not as beneficial as you actually making a public commitment to your plan. If you make the public commitment to the plan, then the technology developers who are creating carbon removal can actually take that to the bank.
LEVITT: And that resonates and multiplies through the system, giving these companies access to capital, which allows them to move much faster and to weed out the ones that are going to work and not work more quickly. Even in absence of a really well-developed final market that might exist by, say, 2050.
ATHEY: But a final market is still going to require regulation because again, nobody really has an intrinsic demand for carbon removal. A company purchases carbon removal as part of a way to meet climate goals, either for public relations or because regulation requires it.
LEVITT: If the U.S. government put a price of $50 per ton for carbon dioxide. And these companies could pull it out of the air for 35, then they could make a really nice profit. aside from any kind of corporate philanthropy. But that’s a long ways off.
ATHEY: Well, sure. And I think that’s what I meant by regulation. So, if you had some type of penalty for polluting, that could be offset by purchasing carbon removal, then that would create a market. But this thing doesn’t exist right now.
ATHEY: But suppose I knew I was going to spend $50 million next year on purchasing carbon removal. But if I know it and nobody else knows it, then it doesn’t really influence financing.
ATHEY: Or if the financers think, there’s 50 percent chance Susan’s going to change her mind and give the money to Ukrainian refugees instead of to carbon removal, then they’re going to count the money as 50 percent. But if I actually commit myself to 50 million next year, then the financiers look at the market size and they decide, yeah, that’s a real $50 million commitment. I can count on that market being there. Therefore, I’m willing to invest into the technology.
LEVITT: And so, these companies got together — Stripe, Alphabet, Shopify, Meta — and they put up almost a billion dollars into this fund.
ATHEY: Yeah, it’s pretty exciting. And a billion dollars, in this market — that was a lot. That was much more than had been spent to date. And it’s really beyond the capacity that currently exists.
LEVITT: You have been both at Microsoft and around big tech a lot and you hear all the time, “Oh, Facebook knows more about us than we know about ourselves. Artificial intelligence algorithms are being used everywhere, and they’re dictating our lives.” As someone pretty close to that space, do you put credence into those fears?
ATHEY: My take on that has actually evolved over time. The value of data really depends on the complexity of the problem that you’re solving. So, I guess the question of “How much do these companies know?” and “How much does it matter that they know it?” really depends on the context. But something that’s changed, I think in the last few years is that a few companies really have figured out how to put together data in ways that makes full use of it. So, I would say like, 10 years ago, even, Google had their search engine, but they had other data sets that were sort of siloed from them and some of them, they had done that by policy, some of it with by agreements to regulators, some concerns about privacy, some just technical considerations. But then over the last, five to seven years, slowly, those datasets have been integrated so that you come up with a super profile on people.
Some things don’t scare me that much. Like, the fact that, you know, people can put together what websites I went to, it’s generally not too dangerous. But if you think about what would happen if there was a breach of a dataset that put together in one place — linked to a single I.D. — every website you went to, every physical location you’ve been to, everything you saw on your browser, everything you saw on your T.V., what was in your watch, what your heart rate was. If you imagine, like, putting together all of these data into super profiles that are really in depth on a person, I think that is pretty risky and there’s a lot more potential for things like blackmail, ransomware, national security issues where, somewhere, someplace in all of that activity, you’ve done something you’re embarrassed of. Those types of things are concerning to me.
LEVITT: Yeah, that makes a lot of sense. because I used to say to people, “Honestly, I’ve been close to some of these firms and they’re not as good at doing things as you think they are.” But I have to say, recently, I’ve come to question that because I don’t go on Instagram very much. I go on occasionally and I buy almost everything that Instagram ever advertises to me. They know me better than I do know myself. I’m like you, I’m not focusing on the negative side. I’m looking at it more from the perspective of a consumer who’s blown away by the fact that Instagram offers me these amazing products which I can’t buy fast enough.
ATHEY: There’s a lot we can do to make all of this safer and more beneficial to consumers. So, I work with startup companies and for a startup company, the thing that gets you funded is having low customer acquisition costs. If you’re acquiring end consumers, even if you need to acquire suppliers — like Uber and Lyft need to hire drivers and they need to acquire customers. So, venture capitalists look at something called the customer acquisition cost, which is just what it sounds like. And one of the main ways that you acquire customers is through advertising. Now, if it gets very expensive to acquire customers on Facebook or Instagram or Google search or YouTube, then you won’t get funded. You have to create more economic value for the consumers than it costs you to acquire them. So, if we have monopolies and market power in advertising that means less startups. So, I think a lot of consumers like all the good stuff they can buy, but they forget that the way that you get all that good stuff is that companies make the good stuff and companies don’t get funded to make the good stuff you want to buy, unless the companies can reach you efficiently.
So, targeted advertising can be really beneficial. It’s beneficial because you’re not seeing ads for things you’re not interested in. It can make it cheaper for companies to find the people who are actually interested in their stuff because you’re not wasting time and space when there’s not a good match. But the thing that I think can make people unhappy is if they think their data’s being used in a way that they’re not comfortable with. So, if companies are requiring their data and they’re somehow using it to limit their choices or to make sure that they don’t have choices in the future — that kind of gives people a lot of discomfort. It’s not so much that you dislike seeing a relevant ad. It’s that you’re uncomfortable with the fact that you feel like you just don’t have a choice about how your data is being used.
LEVITT: Do you remember an interview you did — it must’ve been shortly after you got your Ph.D. — I remember it well, because someone told me about this interview and it was for a teen magazine about how cool it was that you were a Ph.D. economist at M.I.T. and you very much spoke like a mentor to young women. I just remember thinking, “Wow, it’s awesome that the people who publish this teen magazine think that being an economist is cool.” But my colleagues reacted very differently. They were mocking you mercilessly. “Susan’s such a self-promoter.” And I’m sure part of it was jealousy, but no doubt, it was also misogynistic. It was the first time that I recognized a gender bias in economics. And you probably had no idea people were criticizing you in that way at the time.
ATHEY: No, I did.
LEVITT: Oh, you did?
ATHEY: I did. It would get back to me — and it hurt. There were tears over that. I will definitely confess. It made me feel really sad because I guess it felt unfair because I was working so hard. I was in there nights, I was in there weekends. I didn’t know anybody working harder than me. And people couldn’t actually see why I needed to do that. I felt I had been given this privilege. I didn’t choose to be one of the first female economists. There were amazing female economists before me — but I didn’t choose to be lonely there. But I felt — given that privilege, if someone asked me, how could I say no? What kind of person would I be if I was too selfish to take time out of my day to share the privilege I have to try to inspire others? There weren’t many people for Teen Vogue or Vogue or Harper’s Bazaar or any place else to choose from. And that felt to me like not an obligation, but just a responsibility, basically.
LEVITT: I had the Harvard economist, Claudia Golden as a guest on the show. And she’s been in academic economics since the early 1970s, when there were virtually no women in the field. And I asked her whether it had been hard being a woman and she mostly said no. It hadn’t been a big deal. Which shocked me. You haven’t found that being a woman made things easy, have you?
ATHEY: I think it’s a mixed bag. Early on, I got many more opportunities earlier than I think I would have if I had been male. Especially in theory, where I was early, there were just hardly any women. But that also meant that I was working harder with more obligations earlier. So, whether it’s serving on the N.S.F. panel or, being an associate editor earlier. I had more organizing conferences. I organized every conference. I was on every program committee. I was doing, you know, a zillion referee reports, and everything else. And I also felt a lot of pressure because there were a lot of people counting on me to succeed and it would be a very public failure. There was no anonymity there if things didn’t go well. So, I think it was hard from that perspective. You could argue I internalized too much. I took too much on my shoulders. Maybe I should have said no more often. But on the other hand, what I’ve tried to come to lately when I second guess myself or telling other people who are second guessing as well — I don’t know if you want the counterfactual Susan. There’s a counterfactual Susan who’s more chill and doesn’t take on all the responsibility and doesn’t worry about letting people down but she might not also have done the other things that I’ve done. I got where I got to and there might’ve been a more graceful way to get here, but that’s how I got here.
LEVITT: So, you have this reputation for being a superwoman. Is it really true that you fought a coyote and won?
ATHEY: It is really true. Yes. With a broom.
LEVITT: Tell me this story. Why were you fighting a coyote?
ATHEY: I was working at home one day and I heard the chickens squawking, which they do sometimes —
LEVITT: Wait. Which chickens?
ATHEY: Our chickens.
LEVITT: You have chickens at home.
ATHEY: We have chickens. We live on campus at Stanford. So, we have a nice big yard, and we have a chicken coop, and we have chickens that free range around our backyard. They poop a lot, but they make nice eggs. And we had a very sweet pet chicken, Viola. Each batch of chickens we’ve raised from little, tiny baby chicks so we’re very attached to them. And so, I heard the chicken squawking and I looked out the window and I saw a coyote with Viola, our chicken, in his mouth. So, I did what any chicken owner would do, I ran outside and the nearest thing I had handy was a broom, so I threatened the coyote with the broom. Now the coyote did not, in fact, drop the chicken immediately. He was very attached to his chicken, but he did start eventually running away. And then when he got to our fence, he realized he couldn’t make it over the fence with the chicken in his mouth. I whacked at him a few more times and he dropped the chicken and ran away over the fence. My parents grew up in Alabama and Viola would have definitely been chicken soup at that point. But knowing my children were very attached to Viola, I got Viola into a box and Viola went to the vet and got stitches. Then we had an old guinea pig cage, and so Viola came into the house and we nursed her back to health in the guinea pig cage.
LEVITT: Happy ending. So, Viola has been giving you eggs ever since?
ATHEY: Viola is still alive. She’s, I think, seven years old. And she’s still friendly and sweet and kind and laying eggs.
LEVITT: That’s great. I love it. Do you have other superwoman stories that are even above and beyond that one?
ATHEY: I think in some ways these stories can — they can set unrealistic expectations though. It’s hard, I guess, rather than a superwoman image, like I’m still scared when I take on something new. I am completely opposed to letting fear stop me, but I feel it. And if something really matters to me, I certainly power through the exhaustion. But I’m still tired. The thing that really powers me through a lot of challenging situations is just remembering that it’s a privilege to have the opportunity to be able to make a difference. Just that reminder of if there’s a chance I can make a difference, then I should try to do it. There are many problems, we just can’t fix them. There’s nothing we can do. And so, part of the philosophy I’ve taken to things is to say, “There’s so many problems, it’s easy to get overwhelmed by all of them, but, which ones can I make a difference on?” I can make a difference on my own children and I’m going to put the time into that and put the love and the care into that because only I can do that. There are sometimes something like the vaccines where somebody actually wants the opinion of the economist. And if I have to work late or get up early then I’m going to do it.
I like Susan’s sentiment of trying to dispel the superwoman myth. I remember when I was in college, I viewed my professors as gods — wise and all knowing. But then, not too many years later, I had become a professor and I realized I still didn’t know the answers to most of the questions. The intellectual distance between teacher and student is a lot less than most students think. When you hear experts, myself included, speak with great apparent confidence on topics far and wide, remember, there’s a good chance they’re not nearly so certain of the answer as they sound. Well, maybe that’s not true of all supposed experts, but it’s definitely true of me.
Our boxes are packed and we jump on a plane to Germany for a long adventure in just a few days. So, it will be two weeks until you see a new episode. But to give you a little taste of what’s to come, here’s a snippet of my conversation with our next guest. Yuval Noah Harari.
HARARI: We are trying to acquire divine powers of creation and destruction. The Bible starts with God creating animals and plants and humans. We are now in the business of learning how to re-engineer and create animals and plants and humans, and that’s the most important and revolutionary and frightening thing we are doing. Personally, I’m very frightened of the potential of this.
That is historian, philosopher, and author of Sapiens, Yuval Noah Harari. Follow us wherever you listen to podcasts so you don’t miss that episode. And until then, take care. And thanks for listening.
* * *
People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and Freakonomics M.D. All our shows are produced by Stitcher and Renbud Radio. Morgan Levey is our producer and Jasmin Klinger is our engineer. We had help from Alina Kulman. Our staff also includes Neal Carruth, Gabriel Roth, Greg Rippin, Rebecca Lee Douglas, Zack Lapinski, Julie Kanfer, Eleanor Osborne, Ryan Kelley, Emma Tyrrell, Lyric Bowditch, Jacob Clemente, and Stephen Dubner. Our theme music was composed by Luis Guerra. To listen ad-free, subscribe to Stitcher Premium. We can be reached at firstname.lastname@example.org, that’s email@example.com. Thanks for listening.
* * *
ATHEY: Well, I had very long labors, and so I had contractions for a whole day, but I wasn’t making a lot of progress. So, it was at the hospital. I was just hanging out there and yes, I did make a few work-related calls, but that was just keeping myself distracted.
- Susan Athey, professor of economics at Stanford University and a chief economist at the U.S. Department of Justice.
- “Stripe, Alphabet, and Others to Spend Nearly $1 Billion on Carbon Removal,” by Akshat Rathi (Bloomberg, 2022).
- “The Economic Case for Federal Investment in Covid-19 Vaccines and Therapeutics Remains Strong,” by Susan Athey, Rena M. Conti, Matthew Fiedler, Richard G. Frank, and Jonathan Gruber (Brookings, 2022).
- “Advance Market Commitments Worked for Vaccines. They Could Work for Carbon Removal, Too,” by Susan Athey, Rachel Glennerster, Nan Ransohoff, and Christopher Snyder (Politico, 2021).
- “Create a Health Impact Fund to Spur Development and Equitable Use of a Covid-19 Vaccine,” by Aidan Hollis (STAT, 2020).
- “Designing Advance Market Commitments for New Vaccines,” by Michael Kremer, Jonathan D. Levin, and Christopher M. Snyder (NBER Working Paper, 2020).
- “Economist as Engineer,” by Bob Simison (Finance & Development, 2019).
- “The Impact of Targeting Technology on Advertising Markets and Media Competition,” by Susan Athey and Joshua S. Gans (American Economic Review, 2010).
- “Susan Athey, Clark Medalist,” (The American Economic Association, 2007).
- “The Top Draft Pick in Economics; A Professor-to-Be Coveted by Two Dozen Universities,” by Sylvia Nasar (The New York Times, 1995).
- “The Only Covid-19 Book Worth Reading,” by People I (Mostly) Admire (2022).
- “A Playbook for Beating the Next Pandemic,” by Freakonomics, M.D. (2021).
- “Moncef Slaoui: ‘It’s Unfortunate That It Takes a Crisis for This to Happen,’” by People I (Mostly) Admire (2020).
- “Will a Covid-19 Vaccine Change the Future of Medical Research?” by Freakonomics Radio (2020).
- Khan World School Levitt Scholars Program.