Alvin Roth on Market Design, Repugnant Markets, and the Economics of Matching

Alvin Roth

Economists have a much broader view of what a market is than many people do who just read the newspapers and the financial section. And among economists, I probably have an even broader view, because I think about markets that aren't commodity markets and in which prices may not do a lot of work.

So it might help to first think about commodity markets, which is what most people think about when they think about markets. A commodity market is a market like the New York Stock Exchange or the Chicago Board of Trade, where you can deal anonymously with the whole market because all the goods named by particular securities are the same. All the shares of Microsoft you might buy are the same. All the bundles of 5,000 bushels of number two hard red winter wheat are the same. So you don't care who you're dealing with, and prices do all the work.

But in lots of markets, you do care who you're dealing with, and prices don't do all the work. Take labour markets. When you apply for a job, you're not applying for just any job, you're applying for particular jobs. And when you hire someone, you're not hiring anyone, you're hiring a particular person. Those are matching markets, which form relationships. They're also markets, but prices don't do all the work. And there are some markets like that where we don't let prices do any of the work, like putting children in public schools. A given kindergarten class can only hold so many students, and deciding who gets into which class is a market exercise, because markets are the human institutions, the human artifacts, that try to aggregate private information and turn it into collective action.

Alvin Roth

Well, a family is not a market. A corporation is not necessarily a market — it could be a hierarchy. An army is not a market. There are lots of things that aren't markets. Families exchange goods, but they don't do it in a market way. Among other things, markets allow people who don't know each other to engage with each other. But the thing about a family is genetics and lifelong relationships.

Alvin Roth

He certainly talks about information. The phrase he's famous for, which he hardly ever used, is the invisible hand. What he means by that is that collectively we're aggregating our information. We're pursuing our private goals while being constrained by other people's pursuits in ways that might produce good outcomes. That's the joint action of many people, the aggregating of information.

Alvin Roth

Oh, not at all. When you go into a market, you have some information — I have information about what brings you here today, what you're looking to get. Maybe you want to buy wheat from me, or maybe you want to hire me. And part of the labour market's job is to help you decide whether you want to hire me. So we have to exchange information to see whether we're a match — whether I want to work for you, whether you want to hire me.

Alvin Roth

Markets don't always maximise, but they try to reach outcomes as good as possible for the participants, given what everyone else is doing. In labour markets, we sometimes invoke an idea we call a stable matching. It's a market where I may not get the job I want most, and you may not hire the person you want most, but if you hire me, it's because the people you preferred to hire didn't want to work for you — they got jobs they like better. And the people I would have preferred to be hired by didn't want to hire me. So you and I are doing as well as we can.

Alvin Roth

Absolutely. Markets are human artifacts. They're built by people to serve different purposes, and to serve their participants. I'm a market designer. I work in the part of economics that talks about how we should design markets and how we should fix them when they're broken. So the idea that markets somehow appear magically is not a true statement about the way markets are. It's like talking about roadways — roadways don't appear magically, we build them. Language is another human artifact that seems to appear magically. You and I are talking in English, and we can't easily change English, but it changes all the time. We're talking on a podcast — that's a new word we use to describe this thing we're doing. So we're collectively modifying language all the time, even though we don't think of it as an artifact.

Alvin Roth

Society has big influences on markets, but I wouldn't say society designs them. Think about Uber, which is a market for drivers and passengers. It's a company. It was designed by a company called Uber. Cities regulate it — California passes laws about how gig workers have to be treated compared with regular employees. So lots of further design goes on, and Uber modifies its market as it gains experience.

One distinction that's helpful is to talk about marketplaces as small parts of big markets. There's a big market for transportation. Uber is a marketplace in that market. That's mostly what we get to design — we design marketplaces. Uber is a marketplace, Lyft is another marketplace, and municipal taxi services licensed by cities were another marketplace.

Alvin Roth

That's right. And they don't appear in their final form immediately. They get modified by their users, by regulators, by observers who feel impacted by the market even though they're not participants. Think about Airbnb. Airbnb is a marketplace for hosts and travellers. Cities are starting to have some concern about neighbourhoods that might have too much concentration of Airbnbs — that changes the nature of the neighbourhood, and it wasn't covered by existing zoning laws that talked about hotels. So when you look at Airbnb now, it's a combination of the original design, the modifications they put in, the way the users have used it, and the way cities have regulated it.

Alvin Roth

Commodity markets are a great market-design invention. Before the Chicago Board of Trade, if you wanted to buy wheat, you had to inspect it. Every field of wheat is different, so there was a lot of work done before you could buy it — you had to send someone to look at the harvested wheat, measure its water content, see how much hard red wheat was in it. But the Chicago Board of Trade commodified it. By having all these adjectives when we describe the wheat — number two hard red winter wheat — they remove some of the possibilities, so that when you buy 5,000 bushels at a time you don't care who you're dealing with and you don't have to inspect it. If it was farm by farm, you'd care which farm you were dealing with, and you'd want to buy from someone who had reliably kept the soft wheat out of the hard wheat.

Alvin Roth

And there are other podcasts, so not everyone will want to be on yours. Matching markets are markets where you can't just choose what you want. You also have to be chosen. So here we are, you and I.

Alvin Roth

In a commodity market, since you don't care who you're buying from and they don't care who they're selling to, you can just quote a price. You could say, I'd like to buy some Microsoft shares at this price, and if you want to sell at that price, we have a transaction. I don't worry whether you've taken good care of those shares while you had them, and you don't worry what I'm going to do with them. We don't care who we're dealing with. That's the opposite of how you organise a podcast, where you care who you're speaking to.

Alvin Roth

There are lots of individual ways, and some general theories as well. We talk about markets not working well for public goods. If I make some valuable commodity but a byproduct of my manufacturing is that I pollute the air and water, that's often not priced into what I'm selling. I'm creating negative externalities — I'm harming people external to the transaction. That's something markets can't always handle, and we try to fix them with regulations and taxes.

But what I talk about often is not what's traditionally called market failure, but what's called marketplace failure. Think about Uber in a market-design sense. The internet made all sorts of markets possible, like eBay and later Airbnb, but it didn't make Uber possible, because Uber had to wait for smartphones and global positioning satellites — Uber needs to know where you are and where the cars are. As a result, Uber has a fairly simple problem, because they already know what you want: a car to arrive soon. So their job is mostly to match you to the closest driver.

But Airbnb has a very different problem. They can't automatically tell what kind of apartment you want. They have to elicit your preferences. They have to show you pictures of the living room and the view from the bedroom and convince you this would be an apartment you'd like. The form of these markets is very different, and they can fail differently.

Alvin Roth

The title is Moral Economics, and it's about controversial markets. Let me take a step back. I wrote a previous book called Who Gets What and Why, which came out in 2015. That was an optimistic book about market design. It said that lots of markets have failures — they're broken — but sometimes we can fix them, and I talked about markets I'd been involved in redesigning. But it turns out there are lots of markets that aren't working well and are going to be very hard to redesign, because they're controversial. We don't agree on the manner in which they're broken and what we'd like them to do. That's what this new book is about.

Early on I use the phrase repugnant transactions. A repugnant transaction is one that some people would like to engage in and other people object to and think shouldn't be engaged in — not because they're harmed personally, but because they have moral or religious objections.

Alvin Roth

Repugnant transactions are ones that some people want to engage in and others don't like. Disgusting transactions are mostly things that no one wants to engage in. For example, in California, where I live, we have a law making it a felony to sell horse meat for human consumption. You can't buy horse meat in a California restaurant. The same law makes it illegal to sell dogs for human consumption. But there's no law against selling worms for human consumption, nor against selling beverages made of other people's saliva, because those are disgusting — the law of supply and demand takes care of those. No one could run a restaurant that sold saliva-based drinks. On the contrary, there are people who like to eat horse meat. There are parts of the world where it's considered delicious. The reason we have a law against it in California is that we don't like it, collectively, when people eat horse meat.

Alvin Roth

A 1998 referendum. And it's not an old cowboy law from when a horse was a man's best friend. This was a law passed by people who thought of horses as pets. If you tried to be a careful moral philosopher, it would be difficult to come up with a scheme under which eating horse meat should be illegal but eating cow meat is fine. Repugnance is a funny thing — the people who organised the referendum really liked horses, but some of the people who supported it happened to sell beef.

Alvin Roth

Economists have the bad habit of saying, whenever there's a difficult question, "Oh, that's politics." But it's too important to be left to the politicians. There's an old joke that economists study how we make choices and sociologists study how we really don't have any choices. My book is about how, as a society, we try to legislate which choices you're allowed to make and which you're not. And it turns out that to work well — I'm speaking as a market designer — markets need social support. But bans on markets also need social support. We can ban things, but that doesn't mean they vanish. We don't like heroin, and we have lots of laws against it, but there's lots of heroin and lots of overdose deaths every year.

One thing I explore in the book is: how come it's so easy to buy drugs and so hard to hire a hitman? Suppose I said to you, I'm going to come visit where you live, and I'm not going to take heroin on the plane, but you look like you'd know where I could buy some — could you give me some advice? You'd be surprised, you might edit it out of the podcast, but that would be the end of it. But suppose, more privately, I said there's someone in your town who once gave a very bad referee report on one of my papers, and I want him killed — you look like the kind of guy who'd know where I could hire a hitman. You probably don't know, and you wouldn't tell me. And when you mentioned it to your friends, they'd say you should call the police. The police would treat you with great respect and say, call this guy back, tell him to come to this dive bar with $5,000 in twenties and ask for Joe. And then I'd find myself talking to an undercover detective. So as a society we treat trying to hire a hitman very differently from trying to buy or sell drugs, even though our laws about the two are very similar.

Alvin Roth

Legal and moral are different things. But a good way to find controversial markets is to find markets that are legal in some places and illegal in others. A lot of the markets in my book have that property. Another way is to find markets that are illegal almost everywhere but have active black markets. Heroin is a good example — it's illegal almost everywhere to produce and consume, but there's a lot of it. Of the controversial markets I consider, that's the most difficult one, because I don't have good suggestions about what to do. We're losing the war on drugs, but neither does it accept our surrender. In Portland they tried decriminalising opioids, and that turned out to have bad effects on cities and not such good effects on overdose deaths. So we need to experiment. We need to judge our policies not by our intentions, which are pure — we'd like there to be no heroin addicts — but by results. We have to start thinking about how to treat addicts more like patients than like criminals. But we can't just decriminalise without treatment. Maybe Switzerland has a plan that's worked better than ours, and we should be thinking about things like that.

Alvin Roth

Absolutely. One of the first things you learn as a market designer is that the details matter. Uber is different from Airbnb — they have to organise their markets differently. And the solution to drug addiction is going to be different from the solution to whether we should allow in vitro fertilisation, which we do in the United States, but which is not legal everywhere. IVF is a good example. It was awarded a Nobel Prize in 2010 — Robert Edwards — and by then millions of children had already been born through it. But the opposition has to do with the fact that multiple embryos are created and not all are used. If you think the embryo is a living person, you think of IVF as committing a massacre. So people can have really strong opinions, of praise and of condemnation, for transactions that happen millions of times.

Alvin Roth

Let's say morally contested rather than morally questionable, because the people on each side don't think there's any question at all. To start with one that isn't often thought of as a market, but is a matching market: same-sex marriage. Two people want to marry each other and other people think that shouldn't be allowed. It doesn't apparently harm the people who object, because you can't even tell if people are married if they don't tell you — that's why people wear wedding rings. For centuries it wasn't legally recognised anywhere, and now it's recognised in more and more places, but it's still controversial.

Abortion is another. In the United States it was legally banned in most places before the 1970s, but Roe v. Wade made it a right for women to choose whether to terminate a pregnancy, and recently the Dobbs decision reversed that. In his concurring opinion to Dobbs, Justice Thomas said the court should also reconsider its decisions about same-sex marriage, about consensual sex between adults, and about contraception. So some of these controversies we think of as long resolved are still simmering under the surface.

The way I got into controversial markets was thinking about kidney transplantation. Right now in the United States there are half a million people on dialysis, and about 90,000 on the waiting list for a kidney transplant, but we do fewer than 30,000 transplants a year. So most people who could have longer, healthier lives with a transplant will die without getting one. Now, you can get a kidney not just from a deceased donor but from a living donor, because healthy people have two kidneys and can remain healthy with one. My late colleague Gary Becker, who had a Nobel Prize in economics, used to say to me, there's no shortage of kidneys, there's a surplus — everyone has two. The problem is we don't let prices adjust the supply. It's against the law to pay someone to be a kidney donor. So we have about 7,000 living kidney donors a year, but the law requires the kidney to be a gift.

Alvin Roth

The Islamic Republic of Iran. I've read the Shia fatwas, the religious rulings that allow it, and they just say it's not so bad — it saves a life, and saving a life is very important. There's no prehistory of special kidney transplants in Iran. The first kidney transplant is only from the 1950s, so transplantation is a very modern thing. It's not biblical or Quranic. They just didn't see it to be a big problem, and indeed it's not a big problem in Iran.

Alvin Roth

Broadly speaking, two arguments. One is that it's simply wrong. You shouldn't be able to sell body parts, or buy them — that commodifies people in a way we just shouldn't do. The other, more nuanced argument is that only poor people are going to want to sell their kidneys, and they'll be coerced into doing this terrible thing — which is a wonderful thing if they do it altruistically to save someone's life, but a terrible thing if they do it just to be paid. And that it would corrupt the medical profession.

There's a parallel argument for the market for blood and blood plasma. That's different from kidneys — you can give blood or plasma frequently, whereas you only have one extra kidney. But it's also against the law in many parts of the world to pay plasma donors. Plasma is on the World Health Organization's list of essential pharmaceuticals — the products that can only be derived from human plasma. If you wanted to start a hospital anywhere in the world, the WHO says you must have plasma products; you can't operate without them. But they also say, and it's law in many places, that you must get blood and plasma from unpaid donors.

Now, unlike kidneys, there's no widespread shortage of plasma in the world. The reason is that you don't have to pay donors where you live — you can always buy it from the United States, because we export tens of billions of dollars of plasma and plasma products every year, because we pay donors. The people who donate plasma are, on the whole, not as rich as the people who don't — it's something people do to supplement their income. A recent paper looked at what happens when a new plasma facility opens that pays people. Payday loans go down when plasma facilities go up. Some people are living paycheck to paycheck, and donating plasma is a way of supplementing their income that substantively helps them. And it helps people all over the world who have hemophilia or immunological deficiencies. So it's an unusual situation where in many places they think it's immoral to pay plasma donors — and besides, you don't have to, because you can always buy it from the US.

Alvin Roth

Let's come back to paternalism, because we should be of two minds about it. But yes, I think you're right. One of the reasons I wrote the book is that when people make moral arguments, they often don't do the thing economists do all the time, which is think about tradeoffs. We might not like the idea of paying people for kidneys or plasma. But neither would we like lots of people dying of hemophilia or immunodeficiencies. Plasma saves lives. So I personally have no problem with people selling plasma and the US exporting it. We might really hate heroin because we think no one should be addicted — being addicted is like being a slave. But we passed a law saying we'll put sellers in jail for a long time, and it hasn't worked. There was recently a New York Times article about drugs in prisons, where people soak paper with drugs that can then be smoked. The deliveries seemed to be books from Amazon whose pages were full of drugs. If we can't keep drugs out of prisons — surely the place with the most control anywhere — what's the chance we'll keep them off the streets using purely police and military methods? Human beings are resourceful, and this is a market some people want to engage in, so it's very hard to stop them.

Alvin Roth

There are. Canada has had many laws province by province against selling plasma, but has always allowed Canadian Blood Services, the government, to pay. Recently they've started to do so, by allowing a Spanish company, Grifols, which also operates in the United States, to collect plasma in Canada for pay, to be used in Canada. Grifols is a good example, because Spain is one of the countries where it's illegal to pay plasma donors — but Grifols is a subsidiary of a Spanish company that operates in North America and collects plasma here, which is why Spaniards aren't dying of blood shortages.

Alvin Roth

I'm a market designer, remember. If I were asked to design an ethical, legal market in which we could be more generous to donors than we now are, I could do so. That doesn't mean I'm in favour of a market where rich people directly buy kidneys from poor people, pricing poor people out of the market with no safeguards. One thing that would be defensible, but is not about to happen soon, would be to amend the National Organ Transplant Act, which says no one may give valuable consideration for a kidney for transplant, to say — the way Canada has done with blood plasma — that no one but the federal government may do that. Then the kidneys obtained this way would be treated, as deceased-donor kidneys now are, as a national resource allocated on the basis of medical need. So poor people wouldn't be priced out of the market when we had a greater abundance of kidneys.

Right now there's much less ambitious legislation before Congress, a bill called the End Kidney Deaths Act. It would give limited compensation, through income-tax credits, to non-directed donors — people who donate a kidney anonymously without specifying who it should go to. They're very valuable in the kidney-exchange system, because they can spark long chains of donations. We have about 500 non-directed donors a year in the United States. If by being nice to them we could double it, that would be really good for patients, and it would pay for itself, because it's very expensive to take care of kidney patients, and Medicare pays most of the bills. So this seems like a real no-brainer, at least as an experiment. I'm not optimistic, but I'd like to see that legislation passed.

Alvin Roth

Back around 2000 there was a terrible kidney shortage — about 40,000 people on the waiting list when I got involved. Today there are almost 100,000, and many more who would be on the list if being on it were a reliable way to get a transplant. An idea had already been proposed for kidney exchange. It turns out kidneys are a matching market — they're not commodities. You can't take just any kidney; it has to be compatible with your physiology. So you might love someone and be healthy enough to give a kidney, but you can't give it to the person you love who needs one, because they don't match.

One reason people can't get a kidney from someone who loves them is that they've developed antibodies to that person's proteins, so their immune system is ready to attack the kidney if it appears. And one way of developing antibodies is by giving birth to a child — children inherit their proteins half from mum and half from dad, so a mother might develop antibodies to the father's human leukocyte antigens, and then she can't take a kidney from the father of her children, or from the children. So kidney exchange is the idea that maybe you're in that situation, wanting to give to the mother of your children, and I'm in that situation too, and we can't — but maybe I can give to your spouse and you can give to mine. That's a kidney exchange. Over time we've developed it to involve much more complicated transactions, including those chains begun by non-directed donors.

Alvin Roth

Exactly right, and I'd like to see that. Another barrier right now is that if you live in a small country and you're hard to match — you need a kidney but you have a lot of antibodies — it's very hard to get one from a living or deceased donor. It would be nice to do kidney exchange across borders. During COVID I got to go to the United Arab Emirates for the first kidney exchange between the UAE and Israel. That took a lot of doing, but they managed it and have continued. It gives me hope that one day the US and Canada will be able to do kidney exchange together. Canada doesn't have a much bigger population than California, and we wouldn't do kidney exchange in the United States without California. It's a shame that we and Canada can't help each other out — but so far it's just bureaucratic obstacles. There's no moral contest there. They have very different health systems and pay for things differently, and what you need is a year of work by the bureaucrats on both sides to make it happen.

Alvin Roth

Economists are good at thinking about equilibrium, but we're not very good at thinking about equilibration. The process of equilibration is sometimes slower than other things going on in the market, so we may often be looking at markets that aren't in equilibrium. They're moving towards some equilibrium, but the equilibrium is a moving target — as other things change in the economy, the direction of movement changes. That's one of many things we economists don't understand nearly as well as we'd like to. It's a hard problem — a bit like non-equilibrium thermodynamics in physics.

Alvin Roth

It was. First of all, it was terrible employment conditions. Indentured servants voluntarily entered into something like five years of slavery. When I say slavery, an indentured servant who left his job could be captured and brought back — the laws supported that. For a fixed period he was essentially a slave. That's how lots of people bought passage across the Atlantic from Ireland or England to the thirteen colonies. And markets with real information asymmetries can lead to real abuse — presumably they didn't really know what they were signing. You're sitting in Ireland, hungry or ambitious, and some ship captain says you don't have the money for passage, but if you sign this article of indenture, I'll take you across.

So we don't allow that anymore. But sometimes it might work to your advantage to be able to sign an employment contract that says you won't quit. Since that's not a legal labour contract in the United States, we have ways around it. You can agree to work for a company and be paid a significant portion in stock options that only vest after five years. Now you're free to quit anytime, but much of your compensation will stay behind if you leave. That's a workaround — a sign that I might have been willing to hire you at higher pay if you could more credibly commit not to quit.

Alvin Roth

It's easy for lawyers to tell the difference. That's why we have laws and regulations about markets. Regulation is part of market design — regulators play a big role in designing the details of how markets work. Often regulations apply to many markets. The Santa Clara County Board of Health has authority over all the restaurants in my neighbourhood, and that's good, because we all agree restaurants should have clean kitchens without rodents. But when I go to a restaurant, I don't inspect the kitchen, and they don't invite me to. So it's nice that Santa Clara is doing it.

Alvin Roth

You have to look case by case — what are the details? Let's talk about paternalism. Paternalism itself isn't a terrible word. Maybe it should be called parentalism. When we think about children, if you don't supervise them properly you might be guilty of neglect. Behavioural economics is the part of economics that says not all of us are completely good at deciding things in our own best interest. All of us former children who have had enough birthdays to be legal adults sometimes still eat too much or drink too much. So we see proposals like banning giant cups of sugary soda, because once you fill it up you feel you should drink it all. Sometimes those are defeated — Mayor Bloomberg in New York tried such a rule and couldn't, because people said it was paternalistic. But many of us need some protection, and that's what consumer-affairs bureaus are for. There might be financial products that are hard to parse, and the Securities and Exchange Commission should have rules to protect you from pyramid schemes. So the idea that we want to protect vulnerable people — and each other, because we're all vulnerable — is not crazy. There are giant markets for prescription drugs. Prescription drugs are good for somebody, but we don't think you should decide for yourself whether they're good for you; you should have a prescription from your doctor. So we have plenty of markets that aren't laissez-faire. If we talk about decriminalising narcotics, we're not talking about vending machines in elementary schools. You can do something in between laissez-faire and banning a market.

Alvin Roth

Was it Jonathan Swift who wrote A Modest Proposal, in which he proposed eating children — the way to solve a famine is to buy children and eat them? So I'll see your buying children and raise you one: buy them to eat them. Yes, we don't want that. There are lots of laws about adoption, particularly international adoption. Adoption is expensive. If you want to adopt a baby from overseas, you need to jump through many hoops to show you're qualified. You'll need lawyers and international travel, but you can't pay the birth mother for the baby. There's a Hague Convention on that, and it's also illegal in the US to pay the birth mother.

Now, there's a set of transactions that become morally contested when money is added to them. We've talked about kidney donation. Surrogacy is another good one. In California and most of the US, it's now legal to pay someone to bear a child for you, under different regulatory regimes in each state. But in most of Western Europe it's not legal at all — they don't recognise parenthood. And in the British Commonwealth, including Canada, they recognise surrogacy and parenthood but you're not allowed to pay the surrogate. So there's fertility tourism. People who need a surrogate can come to the United States. It's not cheap. They used to go to Ukraine, where it was much cheaper, but the war has destroyed that part of the market.

When you forbid something, you still have to cope with the fact that people might get around it. People have wanted to have children for a long time — we are all descended from unbroken generations of people who had children. So people come to the United States even from countries where surrogacy is illegal, and as a result there are babies. And the German, Spanish, and French courts, where surrogacy is illegal, have to decide who the baby should go home with. A natural answer is, why don't we send her home with her parents? So they've had to allow the surrogate parents to adopt their children.

Alvin Roth

I think I say that in many respects. The laws against surrogacy are designed to protect the vulnerable. When you read the French judicial opinions, they say: we understand there are people who'd like the help of a surrogate to start their family, and people who'd like to help them and bear a baby in return for being paid, but this is wrong, and we don't do wrong things in France. So they're protecting people against their own judgment of a transaction they'd like to do. Nevertheless, there are babies, and there's no one more vulnerable than a baby, so they have to allow adoptions. When we make policies, we have to think not just about our intentions — there should be no heroin addict — but about the consequences. Economists deal in tradeoffs, and a lot of moral contests don't.

Alvin Roth

Medical care is a market. There may be third-party payers, but mostly we purchase medical care. Even where there's a national payer, as in many places but not the US, that doesn't make it not a market — the doctors have to be paid, hospitals have to be built, the equipment maintained. Medical aid in dying is medical aid. People have always had some ability to end their own lives, and the question of medical aid in dying is more often about how to manage your own death. Twelve or maybe thirteen states now allow it — New York State is about to, I think, come August of this year — under restrictive circumstances, and some other countries allow it under less restrictive circumstances. So there's some travel. I start the book with a paragraph about Danny Kahneman's decision to go to Switzerland to end his life, even though he was healthy enough to make the trip. He was still pretty healthy for a ninety-year-old, but in his family he had witnessed some lingering deaths and didn't want his to be that way.

There's a lot of opposition and a lot of interest. Among the strong opponents is the Catholic Church, and the governor of New York is a committed Catholic who spoke publicly about the struggle she had in approving this. What she said is that she was aware of some extremely painful, extended, medicalised deaths, and that it would have been better if New York State had medical aid in dying. In the US it's increasingly common. But we now have two Supreme Court Justices — Gorsuch and Amy Coney Barrett — who wrote, before they were Justices, about their opposition to medical aid in dying. So I wouldn't be surprised if, like abortion, it comes before the Supreme Court again.

Alvin Roth

Certainly we should be more aware. The big lesson, which we've already touched on, is that to work well markets need social support — but so do bans on markets. The example I gave at the beginning: why is it easy to buy drugs but hard to hire a hitman, when our laws are the same? For hitmen it's working. There's very little market for commercial killing. We're successfully deterring hitmen by putting them in jail when we catch them, so let's keep doing it. But the same laws catching drug dealers — more than 40% of our federal prisoners have drug convictions — are filling our jails while we still have plenty of addicts and overdose deaths. So the same laws that work so well against hitmen aren't working well against drugs. I'm not in favour of relaxing the laws against hitmen, but I am in favour of experimenting with ways not just to incarcerate but to treat drug addicts, and to think about how to manage that process. We might learn from countries that have had some modest success.

Think about prohibition and alcohol. In the 1920s and early 30s we had a constitutional amendment that tried to eliminate the production and sale of most alcoholic beverages, and then we re-amended the Constitution to say forget that. Today we still haven't solved the problems of alcohol — it still kills people, through driving under the influence, and it damages families, all the things the prohibitionists worried about. But prohibition didn't prevent those things. There was a lot of alcohol during prohibition, and a lot of crime. The one thing that clearly results from a legal market is that you can no longer buy moonshine whiskey from gangsters — they've been out-competed by the legal market.

Alvin Roth

I think that's a good summary. We can all get on board with that.

Alvin Roth

Thank you.