Should a kidney be sold to the highest bidder? Luigi and Kate debate Nobel-winning economist Al Roth whose algorithm for kidney transplants has saved more than 6000 lives. Roth says matching markets could be used for everything from online dating to the global refugee crisis.
Luigi: How many lives did you save?
Al Roth: Well, so first of all, that’s hard to know, but so far there’ve been about 6,000 transplants arising through kidney exchange in the United States.
Kate: Hi, I’m Kate Waldock from Georgetown University.
Luigi: I’m Luigi Zingales at the University of Chicago.
Kate: You’re listening to Capitalisn’t, a podcast about what’s working in capitalism today.
Luigi: And most importantly, what isn’t.
Kate: On today’s episode, we’re going to interview a Nobel laureate who saved 6,000 lives.
Luigi: Kate, I thought we were running a economics podcast. We’re interviewing a doctor?
Kate: No, we’re interviewing Al Roth, a Nobel laureate in economics who saved at least 6,000 lives through his mechanism to allocate kidneys.
Luigi: Most people outside of economics think that economists only study the behavior of markets, but economists study many things including the way markets should be designed. And not only a traditional market like the stock markets or the oil market, but also markets where there’s not a price.
Kate: When you think of the term markets, you usually think about money exchanging hands, but even in advanced economies, there are markets without money either because people just don’t want to use money or because it’s not allowed. Think about the dating market, for example. Right? People are trying to match up. You’re trying to, in some sense, engage in a transaction that works out for you, but most of the time there’s no money involved even though, I guess, there are dating sites where you can pay, or you can pay a matchmaker, or whatever. But anyway, this isn’t what I’m getting at. On today’s episode, we’re talking about markets where money is explicitly banned. In some cases, these can be what society considers repugnant transactions. Often when it comes to medical practice, such as the exchange of organs, money is explicitly outlawed. Obviously, this makes sense for an organ which you can’t live without—for example, your heart—but actually people have two kidneys, and you can live without a kidney. Yet we’re still not allowed to put a price on a kidney.
Luigi: Now, people might question and say, “Wait a minute. If there’s no price, there is no market. What kind of market are we talking about that hopefully leads to a better allocation?” The challenge in this particular market is there are people who are willing to donate organs. Generally, they try to donate organs to people they care about, but these people are not necessarily compatible, and so you have to find another pair in order to do the exchange. Sometimes you want to do it even more complicated with three pairs of organ/recipient donors.
In a normal market where prices are allowed, if I am a very difficult guy to match, I can start to offer the highest price. This naturally will bring the people to me to try to transact, and this guarantees that the maximum number of possible transactions will take place.
In a market without prices, we tend to be indifferent between two transactions because no consideration is paid. If I swap my kidney with Kate’s boyfriend or if I swap my kidney with my cousin, that is completely indifferent to me. However, because of difference in compatibility, if I swap the kidneys directly with Kate’s boyfriend, I might prevent my cousin to match with somebody else. The complication of the algorithm is to find a way to match pairs that maximizes the number of transplants that can take place. Today Al Roth is going to explain to us how his mechanism saved 6,000 lives.
Kate: Before the innovation that you introduced to the kidney market, can you talk about some of the reasons why it was hard to find a kidney?
Al Roth: Well, the main reason it’s hard to find a kidney is there’s a shortage of transplantable kidneys. This morning, and every morning lately, there are 100,000 people waiting for deceased-donor transplants in the United States, but we only do about 13,000 a year. The main reason it’s hard to get a transplant is there aren’t enough organs. We also can use living donor kidneys because healthy people have two kidneys and can remain healthy with one. There also, there aren’t enough donors, but the additional problem is that even if someone loves you enough to give you a kidney, you might not be able to take their kidney because kidneys have to fit well with your own physiology. That’s where kidney exchange came from. Sometimes you are healthy enough to give someone a kidney, but you can’t give it to the person you love who needs a kidney. I might be in the same situation. I love someone enough to give them a kidney, but can’t give them my kidney, but maybe your patient could use my kidney, and my patient could use your kidney. That’s where kidney exchange comes from.
Luigi: Before we go into the algorithm you devised, in most markets when there’s not enough supply, prices go up, and you have more supply.
Al Roth: Yes.
Luigi: At least that’s what we think in economics. Why this is not a case in this market?
Al Roth: Well, in the United States, and everywhere in the world except in the Islamic Republic of Iran, it’s illegal to pay for a kidney for transplant.
Luigi: That’s interesting, so Iran is more a market economy than the United States?
Al Roth: Yes. Yes, and it’s the only one. You would be less surprised if I said to you, “All of the countries we know of work in a usual way, but the Islamic Republic works differently.” But that’s not the case. One of the things I’ve been led to study through my work in kidney transplantation is what I call repugnant transactions. Transactions that some people would like to engage in, but other people think that they shouldn’t be allowed to. Of course, these are most interesting, most worth studying, if there aren’t obvious negative externalities; we’re all, as economists, we’re already good at understanding markets that some people would like to engage in and other people think they shouldn’t if they have big externalities that harm other people.
With kidneys, that’s not the case or not obviously the case. I think it mostly has to do with worrying about somehow exploiting vulnerable people, the idea that rich people would buy kidneys from poor people and maybe the additional idea that only rich people would be able to get transplants because the price would be driven up. Of course, those are market design issues, but it remains the case that the supply of transplantable organs is much less than the demand and the law requires that the price be zero, that kidneys have to be a gift.
Luigi: So everybody’s making money in the transaction except the donors.
Al Roth: Yes. In fact, the donors may lose money. If you loved me enough to give me a kidney, you would have to incur some expenses. You’d have to come to California probably before the surgeries for some tests. You would have to get a hotel room. You’d have to fly to California, you’d have to get a hotel room a couple of days before your nephrectomy and maybe a couple of days after. You might miss some work. The law actually allows the expenses a donor incurs to be repaid, but it’s very hard in the United States and elsewhere, it’s very hard to repay them.
Luigi: Do you know anything about how in Iran it works? Does it work better in Iran with the price?
Al Roth: It’s a little complicated in Iran. There are some studies of the market. It’s a legal market in Iran, but it’s not one that people are proud to participate in, so there’s some stigma associated with selling a kidney in Iran, which suggests to me that the market isn’t working as well as it should.
An example that the late Gary Becker used to like to give about this was he said how about the transition in the United States from a conscription army to a volunteer army, which of course happened around the end of the Vietnam War. What Gary used to say was that, well, Gary used to say, this is a natural analog for paying kidney donors. What I recall from that time when that debate was going on was there was some concern that American soldiers would come to be regarded as mercenaries because they were going to be paid a wage like a regular job for being soldiers. That hasn’t happened. People are proud to be soldiers. When someone runs for the Senate in the United States, it’s their campaign that emphasizes their military service. When we board airplanes, we’re invited to board after serving soldiers. People are proud to be American soldiers, even though soldiers are paid.
Of course, making something legal isn’t the same as making it unrepugnant. Prostitution is legal in Germany, for example, but I don’t think anyone runs for political office in Germany saying, “You should vote for me because when I was young, I was a sex worker.” It’s still a repugnant market. My sense is that’s going on in Iran with kidney selling. I don’t know a whole lot about that. There are going to be papers quite soon by Mohammad Akbarpour and his colleagues on this subject, but there’s something that although it’s legal hasn’t made it honorable.
Kate: Yeah, I think maybe a closer analogy for the US would be that of plasma donation. Right? A lot of people who need money have to donate their plasma, but you don’t hear people talking about it all the time because it’s not exactly a badge of honor.
Al Roth: Right, and paying for plasma is illegal in lots of places. Canada is debating its laws as we speak. Some of the argument in Canada is that it would be immoral to pay people for plasma and besides, you don’t have to because you can buy all the plasma products you need from the United States. We turn out to be the Saudi Arabia of plasma, and there are-
Kate: Good for us.
Al Roth: ... important medical products like interferon and albumin that are derived from plasma, and we’re giant exporters of those medical products.
Kate: In addition to the requirement that prices have to be zero, are there other constraints legally in the United States to exchanging kidneys?
Al Roth: Well, the main legal issue for living donation is that you can’t pay, so the donation has to be free. Incidentally, of course, kidney transplantation is expensive, as you say. Everyone else in the supply chain is paid. It’s a special program in Medicare. So there are only three entryways into Medicare: you could be over 65, you can be disabled, or you can have kidney failure. Medicare covers the costs of kidney failure for everyone. It’s 7 percent of the Medicare budget. It’s the biggest single program in Medicare. If you have private insurance and you have kidney failure, your insurer will be responsible for you for the first 33 months, after which you become Medicare-eligible. If you don’t have insurance, you’re already Medicare-eligible for kidney failure. It’s a giant American program. Dialysis is the big part of it. Dialysis is much more expensive than transplantation and also not a cure at all. It’s a really good thing for people who have kidney failure and for the American medical system in general, even just financially, it’s a very good thing for someone to get a transplant.
Luigi: Can you explain to our listeners, what is your contribution as an economist in setting up this matching market?
Al Roth: Well, exchange is something that economists study, so the economists, my colleagues and I, so Tayfun Sönmez, and Utku Unver, and Itai Ashlagi, and other of my close analytical colleagues, we certainly didn’t invent kidney exchange. A few were done around the world starting in the 1990s. The first one’s actually in South Korea. There was one in Europe in early 1999 that was met with great repugnance, incidentally. The first kidney exchange in the United States was conducted in 2000, and I had been in a, just a rhetorical way, talking to my class about kidney exchange for some time because I would present a paper by Shapley and Scarf that was published in Volume 1, Number 1 of The Journal of Mathematical Economicsin 1974. They had a model of trading houses with no money, trading a discrete, indivisible good without using money, and my students would say to me, “We use money when we trade houses, you knew that, didn’t you?” I would have to admit that I did know that. I was at the University of Pittsburgh at the time, where a lot of transplants were done, so I started, just as a rhetorical device, saying supposing these were kidneys. You can’t trade kidneys.
When I heard that kidney exchange was starting to be done in the United States, I understood that economists knew some things that would help organize the market. Before we started helping organize the market, there were only four kidney exchanges among the 14 transplant centers in New England. At least one of them had been organized by the patient-donor pairs themselves. Two spouses who were waiting for their spouses in dialysis struck up a conversation, what are you doing here, and one of them said, “My husband, but he has blood type B, and I have blood type A, so otherwise I would give him my kidney.” The other one said, “Oh, you know that’s funny, my wife, she has blood type A, and I have blood type B.” They figured out themselves that they could do a transplant and, and they were able to make it happen in New England. Of course, this is no way to organize a market, so what economists brought to the table was the idea of creating a marketplace that we could have a database of patient-donor pairs and use sensible algorithms to try to propose matches that might go through.
Luigi: Can you give us an idea of a sensible algorithm?
Al Roth: You start with a database of patient-donor pairs, and you need a compatibility matrix. You have to have some idea of which patients can take which kidneys, so which donors and patients are compatible. Mostly, the people enrolled in kidney exchange in the United States are incompatible with their intended donor. You have these patients and donors. The donors are healthy enough to give kidneys, but they can’t give it to their intended recipient. You can start guessing, based on the data about proteins that the donor has, the human leukocyte antigens and the antibodies that the patient has, you can start guessing which donations might go through. That allows you to start looking for exchanges, and you can then run the various algorithms. The first one we proposed was top trading cycles. You can look for cycles of exchange that might go through. Now, we wrote a paper on that and we sent it to all the kidney surgeons we could think of. Only one answered, and that was Frank Delmonico in New England. We helped him form the New England Program for Kidney Exchange.
One of the first things they said to us, is they said, “Look, you have this top trading cycle algorithm. It can propose cycles that are big, but because we do surgeries simultaneously, we can’t do big cycles. To just exchange between two pairs, to do everything simultaneously, you needed to do two nephrectomies and two transplants. You need four operating rooms. You need four surgical teams all available at the same time. That’s it. If you want to help us do kidney exchange, you have to help us organize the market on just exchanges between two pairs,” and that’s how we started. When you do that, you go back to old, elegant results and graph theory about matching and undirected graphs, but of course, you lose a lot of potential transplants. When Jevons wrote about barter, what he wrote about, the difficulty of barter, the reason we don’t see much barter exchange is the difficulty in finding the double coincidence of wants.
When you do exchange between two pairs, they each have to have a kidney that the other one wants. Because some patients are highly sensitized, which means they have trouble finding a kidney that they want, it’s hard to find just that double coincidence, but if you could enlarge it to larger cycles, three pairs, say, where pair one gives a kidney to pair two, who gives a kidney to pair three, who gives to pair one, now you enlarge the possibilities and so forth. Eventually, we managed to convince our surgical colleagues that solving the logistics would be worthwhile, and we could do larger cycles. That helped, and then we were already using integer programming algorithms to find, for instance, the maximum number of transplants you could get from a given population of patient-donor pairs. We don’t actually simply maximize the number of transplants because there’s some concern about the quality of the transplants, but you can think about it as maximizing a weighted sum of the number of transplants.
More recently, we get a lot of benefit from chains that don’t have to be done simultaneously, that are started by a non-directed donor. A non-directed donor is someone who is moved to give his kidney and doesn’t have a particular recipient in mind. You can organize a chain started by such a donor by having the donor give to some patient-donor pair, and then the non-directed donor gives to a patient whose donor continues the chain by giving to another patient-donor pair, and so forth. If you do it so that each pair receives a kidney before they have to give one, if a link breaks there isn’t a tragedy where some pair has had a surgery that didn’t help them and no longer has a kidney with which to participate in kidney exchange.
Some chains have been very long. For a time at least, I don’t have the latest data, the average chain in the United States had five transplants in it, so 10 people in the picture, five donors and five recipients. We get really a lot of use out of chains. More than half of the kidney exchanges that go on in the United States now come through chains. Kidney exchange in the United States accounts for 10 percent to 15 percent of the living donor transplants in the United States. On the one hand, that’s a great success. On the other hand, these victories I’m telling you about are victories in a war that we’re losing because there’s a diabetes epidemic. There’s other things going on, so the number of people in need of kidney transplants is growing.
Luigi: What are the major obstacles to setting up bigger and bigger exchanges everywhere?
Al Roth: Well, there are lots of obstacles. I don’t understand them all. Some of them surely have to do with incentives. Dialysis is a very big business, and many of the people who could profit from transplantation are in dialysis. The law requires dialysis centers to inform patients about the possibility of transplantation, but it seems possible that they do so in a way that isn’t effective in encouraging patients to seek out transplantation. I think one of the big obstacles is reaching people who have kidney failure early in the course of their disease and letting them know that transplantation is a possibility.
Another one is letting them know that kidney exchange is a possibility because I think what can often happen is you need a transplant, maybe you don’t know about living-donor transplants, or maybe you do but the people in your family are incompatible, or have kidneys that are compromised themselves. A lot of kidney disease has a genetic component so that someone who has kidney failure, a lot of their relatives might have kidney failure too. I think that if people, early on in their diagnosis, were aware that transplantation, and in particular kidney exchange, were possible, then we’d see more living-donor transplants. Again, Gary Becker used to say, “There’s no shortage of kidneys, there’s a surplus because you have two.”
Kate: Based on your research, what real-world problem do you think needs solving the most?
Al Roth: Well, there are lots of matching markets, markets where you can’t just choose what you want, but also have to be chosen, that aren’t working very well. Of course, the ones that are working worst are not necessarily the most ripe for redesign. Sometimes the reason a market is working badly persists in keeping it working badly, but one matching market that’s working very poorly right now is refugee resettlement that we’re seeing lots of big migrations of people and we’re not good at resettling them, but neither are we good at keeping them where we would like to keep them. It’s a matching market.
You can’t tell refugees to stay at home. Neither can they choose where to go. When you see people putting their children in small boats in the Mediterranean, it’s clear that there’s something wrong with the way that market is working. But of course, it faces all these very serious issues. Lots of people, well lately, there’s been lots of anti-immigrant feeling around the world, not just in Europe, but also in the United States. If we think that there’s going to be large-scale human migration in the future, which there might be in the coming century, for example if the sea levels rise, then we’d better learn from our current failures how to do it better. I would think that that’s something that we should be very actively thinking about.
Kate: Who do you envision on the other side of that match? Is it a host family that’s interested in taking in a refugee family? Is it a municipality that’s interested in piloting a program, or is it a state?
Al Roth: I think that that’s a great question. There’s matching that happens at each moment. The easiest part to tackle, and the part that market-design economists have started to tackle, is how do you place, and house, and take care of refugees who have already been granted asylum in your country. Once refugees are granted asylum, you have to settle them. There was a lot of sentiment among refugee resettlement authorities that refugees should be spread thinly around the country with the hope that that would make them assimilate easier. Of course, when we observe refugee communities, that’s not how they like to be settled. The example I like to give is that there’s a big Somali-American community in Maine. It’s not because of the cross-country skiing that attracted them. It’s that there started to be a Somali community there, and it turns out if you’re a new Somali migrant to the United States, that’s a nice place to go where there are people who can help you out, get started. We have to start thinking of not randomly allocating refugees but thinking what is needed to help them integrate into the host economy.
Then there’s the harder question of who should be granted asylum in which countries. Again, when you see people in little boats, it suggests that we’re not doing a good job, either of the deterring them from coming or of placing them effectively. I say little boats, I’m thinking of Europe and the Mediterranean. In the US, we see unaccompanied children showing up at the border. The system is working very badly. We are not in control of it. That’s something that I would like to see us do a better job of.
Luigi: We have talked about organs. We have talked about immigrants. Can we talk about a lighter application of this, because also the dating market is a matching problem, right?
Al Roth: Absolutely.
Luigi: And generally we don’t use prices for that.
Kate: I have written down, “What do you think about dating apps?”
Al Roth: Right. Well, dating apps are great to talk about because, of course, they’re matching markets and they run into the same sort of generic problems that all marketplaces run into. First, to be a successful dating app, to be a successful marketplace, first you have to make the market thick. You have to get lots of men and women or people who want to date each other on the site. Then you run into congestion problems if you ... Some of the early dating sites, there are so many men and so many women that the internet traffic in which they tried to meet each other starts to be spammy. People with attractive profiles get more emails than they can answer so people who find their emails aren’t being answered send more and more superficial emails to more people, and that gets you a bad equilibrium. A new generation of dating sites came up that tried to curate the email to say that only women could initiate a conversation, for example, or to get rid of conversations entirely and just have people swiping left and right. Those are efforts to deal with congestion.
There’s the whole question of safety and reliability, of should you curate the people on the site so that you aren’t afraid of meeting them. Nevertheless, Americans seem to meet each other, in increasing numbers, through internet dating sites that result in marriages. I think that that has to do with the fact that other matching markets have become less thick. Right? There was a time when few Americans went to college and it made a lot of sense to marry your high school sweetheart because that was the moment in your life when you knew a lot of people of marriageable age, of your age, who were single, and that would be a good time to get married because afterward it might be hard.
Increasing enrollment in colleges meant that wasn’t the last thick market you were going to see. Increasing female participation in the labor force meant that all of a sudden we had people graduating from college and going into work environments where the marriage market might not be so thick. It’s not always so easy to marry the person in the next cubicle. I think part of the growth in internet dating sites is related to the rising age of first marriage and is an effort to keep the market thick, to give people opportunities when they don’t have them day-to-day. I imagine that that’s going to remain a important part of courtship.
Kate: Thank you so much, Al, for joining us.
Al Roth: Well, thanks for having me.
Luigi: Al is a great guy. His contribution is extremely important for economics but more importantly, is also very important for humankind. However, listening to his discussion and listening to how complicated it is to do these matches even with his algorithm and how many people are left out, the question arises: Why don’t we pay for people to donate an organ? Of course, not your heart because it means that you are killing yourself, but what about a kidney? People can very happily live without a kidney. If they’re willing to donate a kidney for money, why is it so wrong?
Kate: Luigi, how much money would you have to be paid to sell your kidney?
Luigi: Wow, that’s a good question because I would give a kidney to my wife and my kids, but I don’t think I would sell it for money.
Kate: All right. Fair enough. I’m not sure there’s a price that I would accept either.