An Interview with Alvin Roth, 2012 Nobel Prize Economist
Alvin Roth is the Craig and Susan McCaw Professor of Economics at Stanford University, the Gund Professor of Economics and Business Administration Emeritus at Harvard University, and the 2017 President of the American Economics Association. Mr. Roth was awarded the 2012 Nobel Memorial Prize in Economics for his work in the field of applied game theory, and more specifically, the practice of market design. He has developed programs for matching doctors with hospitals, school children with schools, and organ donors with patients. He is the author of Who Gets What and Why: The New Economics of Matching and Market Design.
The Politic: First off, could you give a brief outline of ‘market design’ and its relation to the idea of a free market?
Alvin Roth: There’s a romantic idea of a free market that doesn’t bear much scientific examination, which is that markets somehow happen without human intervention. I think that’s a poetic idea but not true and has never been true. Markets are human artifacts; we built them. Just to take contemporary examples, nobody thinks Uber just emerged, or Airbnb. Those are companies. They run marketplaces.
When the first people who made stone axes started noticing that other people were interested in them, they set up some kind of marketplace. They were prepared to trade the stone axes. So markets do emerge, but they emerge through human activity. They’re tools. They’re artifacts. So market design is studying that process consciously, and thinking about how we can improve it.
What would you say are the constituent features of markets?
One thing about markets is you have to bring the participants together. In my book, I talk about making the market ‘thick.’ One way to make the market thick is to coordinate on a time when people will come together to trade. So ‘unravelling’ is about how that can sometimes break down. Timing gets diffuse and the market stops being thick. That’s from the point of view of starting with a thick market and then losing the thickness.
But, of course, you have the opposite problem when you’re just starting. If you have a farm stand and you sell your tomatoes when they’re ripe, then you’re selling to people who come along the road. But if you and the other farmers get together and have a farmers’ market every Monday, then you have a thicker market. So we see market design like that.
We have a bunch of farmers markets around here. The one we go to is on a Sunday morning at California Avenue in Palo Alto. That’s a designed market. The city of Palo Alto has to close the street. The sellers are always in the same place– they reserve spots. There’s obviously some design underlying the market, so that when I show up and want to buy raspberries, they’re there, and I’m there. That’s making the market thick.
It’s also a market that’s more congested than the adjacent supermarket. Supermarkets are a great invention too. In the supermarket, the stuff has been on the shelves longer, but you push your cart around, and you only pay once. The supermarket solves some of the transaction costs, some of the congestion problem of buying a lot of stuff, since you only have to pay once.
At the farmers’ market, you pay for the raspberries, and then you go somewhere else and pay for the strawberries, and then you go somewhere else and get breakfast, and then you pay for that. So it has some advantages over the supermarket, and it has some disadvantages. It’s a different design. Market design is about understanding how the rules of the marketplace change performance, and how we might help markets perform better if we see markets that are having difficulties.
What are some domains we might not conventionally consider markets, but that you think we should?
I think of a lot of things as markets. When we conventionally consider markets, we often focus on commodity markets. Commodity markets deal in commodities, by which we mean things that can be traded anonymously; you don’t have to examine them closely because they’re all the same.
Now, there may need to be a standard-setting institution to make sure the commodities are all the same, that enough No. 2 Hard Red Winter Wheat is actually No. 2, so that you wouldn’t have to call it No. 3 Hard Red Winter Wheat, and that enough of it is red and not white wheat. The Chicago Board of Trade makes rules about what makes No. 2 Hard Red Winter Wheat.
There’s an Ethiopia commodity exchange, which among other things, trades coffee. It used to be if you wanted to buy coffee, you had to send a man to Addis, and he had to taste it. He would put a cup into a bag of coffee, cup the beans that came from the middle, and form an impression of that bag of coffee. He might have to examine each bag.
But now they have anonymous tasting. And why anonymous? Because you wouldn’t want tasters to be able to be influenced by the farmers to give them a higher grade than they deserve. What that means is you can order coffee – if you’re a coffee chain – from Addis without sending a man there. You say, “I want to buy a lot of coffee.” Then you buy it at the market price. You don’t have to taste it bag by bag, so it’s made that coffee a lot more accessible to the global market. That’s design, and that’s what we normally think of as markets.
The things we don’t normally think of as markets– I’ve concentrated a lot on markets where prices do very little work like kidney exchange and school choice. But I think the prototypical markets are markets like labor markets, where prices do some of the work but not all of it. I teach a lot more economics courses than I would if I weren’t paid to do it. But prices hardly do all the work of deciding who gets what. Labor markets are matching markets: markets where you care about who you’re dealing with, and where you have to be chosen; you can’t just choose, and prices don’t decide everything. In my work, the markets that most surprise people that they’re markets, are ones that don’t use prices at all, like kidney exchange.
A Kidney exchange seems like a very specific, tangible example of something we might not consider a market. But you’ve also written about obscenity, profanity, and blasphemy. How would you categorize those areas that might seem a bit more intangible?
When I talk about repugnant transactions, I talk about transactions that some people try to prevent entirely. Blasphemy, to many Americans, seems like a concern that comes from a distant past. But in Pakistan today, blasphemy, and blasphemy laws, and popular reaction to what is seen as blasphemy, are very real issues. People live and die on these issues.
A repugnant transaction is a transaction that some people would like to engage in, and others don’t think you should be allowed to engage in at all. And with blasphemy, those are religious transactions. There are people who don’t think you should be able to worship in certain ways. And we have wars where people kill each other, still, in the 21st century. So human beings have a lot of concerns about how other people behave, even when it’s not always clear why we should have these concerns. That’s a fact of being human that I would like to understand better, and one of the reasons why I study repugnant transactions.
What would be your basic method for identifying, on a case by case basis, high-impact areas for market design? For instance, why did you personally choose to focus on school choice and kidney transactions rather than profanity and blasphemy?
I’m interested in all sorts of things. I’m trying to understand repugnance better, and that gets me into things like gambling and profanity. But I chose the things I worked on for a bunch of reasons. Partly it was that these were things I thought I could help improve. Organ transplantation is a set of transactions that are not working so well; we have many more people who need transplants than we have transplants available. Kidney exchange is a way of making transplants more available. And school choice and things like that are similarly things that if they work better, they make families better off.
But part of how I was drawn to them was because of the mathematics. These happened to be things that it appeared, having studied the things I studied, that I’d be able to help with. There are other pressing problems in the world, some of them I pay attention to and look at from a distance but where I have no personal expertise that I can offer to help intervene. I’m interested that blasphemy is a big deal in parts of the world, but I have no helpful suggestions about how to coordinate those concerns with the concerns of other people to worship differently than the majority in the country where they live. I like the conclusion we’ve mostly reached in the United States that religion should be a personal matter, but I’ve noticed that it’s not everywhere.
In your article, “Repugnance as a Constraint on Markets,” you write that “almost whenever [you’ve] been involved in practical market design, the question of whether certain kinds of transactions may be inappropriate has come up for discussion.”
How do you approach issues that may be ethically constrained? Do you think there’s a way to minimize the value judgements you make as an economist?
It’s interesting you ask the question that way. For a long time, economists seemed to agree with the implicit assumption of your question that we should minimize the ethical judgements we make. I’m not sure that’s a universal position anymore. Among economists, there’s some discussion that maybe we should be taking ethical positions about some of the markets we see, and especially with repugnant markets.
Remember, when I use the word ‘repugnant,’ I use it as a technical term: I say a transaction is repugnant if some people would like to engage in it and other people think they shouldn’t be able to at all. I don’t use repugnant as a synonym for ‘disgusting,’ which means “I don’t like it,” or maybe “nobody likes it.” So these are markets where there’s a difference of opinion. And the question is, should we be taking ethical stances? It’s possible that we should, but there’s also a lot of practical questions that I think get missed in the ethical discussion.
For instance, we might all agree that crack cocaine is not good for anyone, and no one should ever consume it. We would never advise people to consume it. We wish no one ever consumed it. So we have laws against narcotics, including crack cocaine. But we also have to notice that although we have laws saying no one should buy or sell crack cocaine, it’s widely available in many parts of the United States. So we don’t like it, partly for moral and ethical reasons, but at the same time, we haven’t been able to make it go away. As economists, I think we have to think about that too.
That’s one of the ways in which I’m thinking about black markets. We contribute to the design of the black market for cocaine by the ways we pass laws against it. There are different ways to pass laws against it. In Portugal, drug addiction has basically been decriminalized. The selling of drugs is still a crime, but having drugs for your personal use is not a crime; it’s regarded more as a medical condition. You are an addict, that’s a diagnosis. How should we treat that?
That’s a different story than criminalizing addiction, with the treatment of choice being incarceration. Indeed, half of our federal prisoners have drug convictions. We work hard to treat drug addiction through incarceration, but we have 60,000 opioid overdose deaths a year right now, so maybe we should be reconsidering the market design. That is, we tried to design a market in which no one used drugs, it looks like we are not succeeding, and there are consequences to that. So, maybe that market isn’t working well, and we should be considering a redesign to reduce the harms that the present black markets are causing.
Do you think we’re seeing a shift toward finding drug crimes less repugnant?
I think we find them really repugnant, but we haven’t been able to make them go away. Some things we hate and we can make them go away and that’s great. But not only do we have a lot of drugs; we also have a criminal enterprise and a medical catastrophe. The people who have this thing which we might call a ‘disease,’ a drug addiction, are not getting treatment that they would if it were a disease like tuberculosis. So when markets don’t seem to be working, we should reconsider how we redesign them. That’s the general idea of market design. And of course, these kinds of markets and black markets didn’t just evolve, they’re partly shaped by all our legislation. Cocaine and heroin are Schedule I drugs, and strangely enough, marijuana remains a Schedule I drug in federal law. (Schedule I means drugs for which there’s no medical use.)
But on the other hand, I live in California where the most recent referenda moved marijuana from being available by prescription only for medical purposes, to being available for recreational purposes. And a bunch of states, not just on the West Coast, now make marijuana legal, while others make it illegal.
In the state of Idaho, marijuana is illegal. It’s not legal for recreation, and it’s not legal for medicine–it’s just an illegal Schedule I drug. On the other hand, if you step across the border into the state of Washington, it’s legal. So I anticipate that, maybe in the near future with a different presidential administration, we’re going to see marijuana become legal everywhere, just because it’s really hard to enforce or to want to enforce real criminal laws against marijuana in Idaho, when right across the border, people are buying marijuana not from criminals, but from bakeries where you can order edibles. So markets have some reach: it’s going to be hard to make marijuana completely illegal in Idaho when it’s completely legal in Washington.
We’ve seen that the illegality of sports gambling, an activity some people might consider repugnant, was recently overturned by the Supreme Court. We’re also seeing a trend towards decriminalization of both prostitution and drugs. So why not organ selling?
We have concerns about the unregulated sale of organs, that it might turn into something we don’t like, or that it might victimize poor and vulnerable people. When we talk about living people, we mostly talk about kidneys. They’re sort of special– you have two, you can remain healthy with one, you can give one to someone. You can also sell one to someone, but that’s illegal just about everywhere in the world. One reason it’s illegal is we worry that somehow rich people would buy kidneys from poor people, poor people would get priced out of the market, and donors would be taken advantage of.
And in the black markets for kidneys, all of those things are true. It’s against the law to buy and sell kidneys, but there are black markets. I know of some in far corners of the world. And we don’t like the way they work. But of course, we don’t like the way black markets work for drugs either. During Prohibition, when we had black markets for alcohol, we didn’t like the way they worked, and that’s why we ended Prohibition. And of course, the legal market for alcohol looks nothing like the black market for alcohol.
Al Capone, the famous Chicago gangster, was a Prohibition Era figure. His criminal organization was a creature of Prohibition. But today in California, if you go into a nice wine store, they’ll sell you very expensively aged whiskey and fine wine. Yet it’s not at all like buying rotgut moonshine whiskey from gangsters. So you can start thinking about how we could increase the supply of organs by allowing some compensation in a regulated way, or at least by removing some of the financial costs of being a donor.
As we speak, a Pennsylvania Congressman named Matt Cartwright is getting ready to introduce a bill – he’s a Democrat – that will have bipartisan support when it comes to the floor. It’s called the Organ Donation Clarification Act. It says that we never intended to make a little bit of experimentation with non-cash incentives against the law, so let’s be clear that testing ways to increase the organ supply is not illegal. This would be an amendment to the National Organ Transplant Act. The idea is that maybe we could offer, for instance, lifetime health insurance and good care to prospective donors. His amendment is still getting wordsmithed. But imagine that we treated kidney donors the way we treat veterans, so that you could go to a veteran’s administration hospital when you’re ill because you served your country and put yourself a little bit at risk and saved someone’s life by donating a kidney.
I think that might pass the repugnance test, that is, that we could convince our fellow citizens that things like that shouldn’t be against the law given the harm that comes to us from the grave shortage of organs for people that need them. So repugnance can change.
That’s basically your idea that x + $ can make x repugnant, even if x wouldn’t be considered as such otherwise, right?
That’s often the case with repugnant transactions. Think about prostitution. Prostitution is illegal almost everywhere in the United States with the exception of a couple of counties in Nevada. But sex outside of marriage is no longer illegal. It was in the time of the Puritans, but it isn’t now. So we’re not repulsed by the idea that an unmarried person might have sex, or that even a married person might have sex with someone he or she is not married to (although that raises lots of other issues). So we’re not repulsed by sex, but we are repulsed by prostitution– and that’ sex plus money. It’s adding money to it that makes it repugnant.
And of course, betting on sports or paying college athletes is somewhat similar. Those athletes aren’t paid. At Harvard and Yale that might not matter much, but here at Stanford, some of the graduates go on to play professional sports. They get paid quite a bit when they graduate, and arguably they are valuable athletes even before they graduate. But they can’t be paid. How do we feel about that? We admire competitive athletes, but we put up barriers to them being paid while they are students.
Actually, if you Google my name, Alvin Roth, and do it a little carefully, you will find a 1930’s basketball player for City College who was convicted of a gambling offense, of fixing a game I think. It wasn’t me!
What do you think of the role of paternalism in repugnance?
Paternalism is a very interesting subject, and I have colleagues who are actively looking into it now– experimental economists like Muriel Niederle. Paternalism is pretty interesting, partly because if I say to you, “you’re being paternalistic to me,” I’m sounding offended. But if you fail to be paternalistic to your children, you might be accused of neglect. In other words, paternalism is your obligation to your children– you can’t let them play in the street, you should make them go to bed at a reasonable hour, you should help them eat healthy food. As I can tell you having raised children, and now having grandchildren, children don’t always agree with their parents in where their best interests lie, and it’s parents’ responsibility to be paternalistic. ‘Parentalistic’ might be a better word. What happens when you turn eighteen that makes paternalism non-agreeable? Part of it has to do with liberty– we don’t want people telling us what to do.
On the other hand, we enforce lots of laws that are meant to protect us in what you might think of as a paternalistic way. Think of the Food and Drug Administration. There are drugs available in some parts of the world freely that are only prescription drugs in the United States, or might not be available at all in the United States because they haven’t passed FDA safety tests that might or might not be too stringent. There are plusses and minuses to saying a drug can’t be sold. It protects people from inadvertent harm, but it also protects people from the good effects that the drug might have. So that’s paternalistic. We let the FDA decide what drugs are available for us.
Behavioral economics is all about how ordinarily competent adults – real human beings – might not be ideally able to protect themselves against every harm. Maybe we could use some advice. Maybe we should have Social Security so that people aren’t destitute in their retirement. Those are paternalistic, but by and large we have Social Security, and we approve of it– good paternalism. But not everyone does; there’s a political divide. We have a Libertarian political community in the U.S. that might disapprove of Social Security.
So paternalism is a loaded word. It’s a part of life that, of course, is a very good thing when you’re raising children. You’re supposed to take care of your children and protect them from harm when their judgement isn’t well-developed yet. Of course, behavioral economics tell us that sometimes we may not have good judgement. We have consumer protection laws that are designed to protect you against certain kinds of misleading advertisements, and things like that. Those are a little bit paternalistic.
But of course, some paternalism we resent. New York City tried to ban extra-large soft drinks because it’s hard not to finish your food. When your parents were being paternalistic to you, they told you to finish whatever was on your plate. So you’re good at finishing what’s on your plate, and if you get a super big soda, even when you’re not that thirsty anymore, you might feel like you ought to finish it, and that might be bad for you. The question is whether the government should intervene, or is that something where we should just say to you, “don’t order such a big soda,” or “throw it away when you’re not thirsty anymore.” We’re divided about that.
We’re okay with poorer workers taking higher-risk jobs in return for wages, but we’re not okay with them selling their organs for money. What do you think is the fundamental difference between these two cases– why do you think we consider it paternalistic in one case and not the other?
I don’t know, I wish I understood repugnance better than I do. Those are puzzles. We had some discussions like this when I was your age. The Vietnam war was on, and we had a conscription army. Between then and now, we moved to a volunteer army. Back then we had this discussion about whether being an American military person would still be an honorable profession when we no longer had conscription, or whether we would come to regard American soldiers as mercenaries. And of course, we can argue about whether the other aspects of the transition were successful, but we successfully made the transition to a paid, volunteer army without losing respect for our soldiers.
When I go on an airplane, before everyone else boards, they invite serving soldiers to board. So we honor our soldiers even though they’re paid. No one says, “you’re paid, you’re a prostitute.” To the contrary, if someone’s running for Senate and he was an American serviceman, that’s a big part of his campaign: “vote for me– when I was young, I was a marine, a trustworthy and patriotic person.”
Now, in Germany and much of Europe, prostitution is legal. But that hasn’t made it un-repugnant. No one runs for political office in Germany saying, “vote for me– when I was young I was a sex worker.” That’s a legal thing to say, a legal job to have had. But if you were running for parliament in Germany and you had been a sex worker, it would be your opponent who brought this up, not you. Unlike if you’re a soldier. So being an American soldier has not become repugnant by adding money to it, by paying real wages instead of conscripting people. And being a prostitute did not become un-repugnant by making it legal.
And the reason I draw this analogy is because soldiers are different from prostitutes – not because they’re the same – but in both cases, we pay soldiers that’s how we get a volunteer army, and in Germany, it’s legal to be a prostitute or to pay for sex, but being legal or not legal does not determine whether something is repugnant or not, although it reflects that– it’s correlated with that.
So I don’t understand repugnance. I would love to understand it more. I think economists should understand repugnance more because markets need social support, and understanding which kinds of transactions gather social support, and which don’t, is something that is too important to be left to the philosophers. Economists should be investigating that in the empirical and scientific ways that we investigate other economic phenomena. But we have spent too little effort on that so far, in my opinion.
Final thought: jumping back to your position on the role of economists in questions of morality, I’m not sure your colleague Thomas Sowell would agree!
I think there’s a lot to be said for technocratic advice. I should be able to sit at the table with people who might have different ethical judgements, different religious beliefs, and talk about what will happen if we do certain things. That’s why I spoke about crack cocaine. Because I think most people agree that we’d like not to have that for sale in the United States, but the fact that it’s for sale anyway, even though we all agree on that and we’ve made laws against it, is an important fact. You can imagine discussing that as an economist without in any way suggesting that I approve of using crack cocaine– you shouldn’t use it even if it’s legal. So there’s a very useful technocratic conversation that economists can engage in that doesn’t necessarily have any ethical content. It doesn’t matter whether we like crack cocaine or not, our plan for having none hasn’t worked, what should we do? But at the same time, when you talk about sex work for example, the literature on sex work is actually quite divided among two parts that don’t talk to each other so much.
One part talks about human trafficking– slavery, which we all abhor. The other part talks about labor, that there are sex workers– that’s their job, they would like it not to be an illegal job, not to be a dangerous job – and by making it a crime, we make their job more dangerous, and why are we dishonoring them when they are living in ways that are more profitable than alternatives available to them. Both of those are sensible literatures that have ethical content to them, but also have empirical content to them. When we talk about sex workers, what’s the composition of volunteer sex workers who are being criminalized because we disapprove of them, compared to trafficked children and people in various ways who are not engaged in voluntary labor, who are effectively enslaved?
I guess any discussion of sex work needs some empirical infusion. My inclination is that trafficking is like crack cocaine. We don’t approve if it in any circumstances, but if there’s a lot of it, we have to think better about why there’s a lot of it, and what we should be doing if what we’re doing now is failing. And on the other hand, if sex work is largely voluntary, then what can we do to make it safer for the people who engage in it? It’s illegal almost everywhere in the United States, but we haven’t gotten rid of it, so in that respect it’s similar to crack cocaine. It’s a black market that we haven’t been able to stamp out for millennia. It’s called the world’s oldest profession.