Leo Katz on Why the Law is So Perverse
Aug 29 2016

katz.jpg Leo Katz, professor of law at the University of Pennsylvania, talks with EconTalk host Russ Roberts about his book, Why the Law Is So Perverse. Katz argues that certain seemingly inexplicable features of the law are the result of conflicts between multiple objectives that the law or the courts must trade off against each other. Katz also argues that structure of the law and how it is enforced are analogous to certain inevitable ambiguities of collective choice and voting theory.

Michael Munger on Choosing in Groups
Michael Munger of Duke University talks with EconTalk host Russ Roberts about his latest book (co-authored with Kevin Munger), Choosing in Groups. Munger lays out the challenges of group decision-making and the challenges of agreeing on constitutions or voting rules...
Dan Klein on Knowledge and Coordination
Dan Klein of George Mason University talks with EconTalk host Russ Roberts about the ideas in Klein's new book, Knowledge and Coordination. Klein argues that allegory is a powerful way to think about outcomes of emergent order. He goes deeply...
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


Aug 29 2016 at 1:08pm

Just want to thank Russ Roberts and Leo Katz for such an excellent and thought-provoking podcast.

I’ve been listening to EconTalk weekly since 2009 and this has been one of my favorites.

I also had a lightbulb moment when Russ pointed out that maybe the reason we as a society are such big fans of third-party health payments is because we do not want to force people to make life and death decisions with money in mind. Great observation.

As for the prisoners consenting to torture, isn’t that a bit like Singapore? As far as I know, criminals in Singapore are caned instead of fined or imprisoned. It certainly is egalitarian: money has less relative worth to the wealthy, so fining is unfair punishment, and some people’s time is worth way more than others’ — but everyone feels the same physical pain, presumably. However, people absolutely flip out about how evil Singapore is. It’s very interesting and I think Leo Katz has nailed it: some things you just cannot consent to.

Additionally, there is the case of that German cannibal who found (and ate) a willing victim and his murder conviction seem to back up Katz’ argument.

Aug 29 2016 at 3:02pm

I’m surprised the subjective theory of value wasn’t mentioned. This talk provided several examples of how value is subjective, value rankings are only ordinal, and interpersonal comparisons of subjective utility and aggregations of subjective utility are not possible (do not provide meaningful results). The further point raised is that we are sometimes uncomfortable allowing people to make their own choices, and/or we intuitively feel certain norms or certain domain experts should get to decide on behalf of individuals (“for their own good”). I believe these examples don’t have to be viewed as problems unless we’re coercing people. If free people of sound mind are making their own choices, we might not always like the result, but we’ve addressed the ambiguity and done so in a way that preserves a persons right to life, liberty, and the pursuit of happiness on their own terms. I think the big issues are accepting the subjectivity of value, and making people free from coercion (free to make their own choices). The alternative is some form of elite supremacy like government where individual don’t have rights, but only privileges that can be revoked by the rulers (whether dictatorship, majority rule, or whatever).

Aug 30 2016 at 2:28pm

Great episode, I may have to read the book.

Early on in the episode I was reminded of another theory of law that EconTalk introduced me to: Coase described laws as necessary to reduce transaction costs in what would otherwise be a world where bargaining alone could efficiently resolve all conflicts. Many of the hypotheticals that Katz introduced struck me as situations where the cost of having to deliberate would outweigh the gains of the few who could choose to go against the norm e.g. a doctor spending time on presenting a patient the option of giving up their legs for another’s hand is generally wasteful, and so per Coase it is reasonable that the choice is not presented.

Later in the episode it I realised that Katz was making a different point, that the law may intervene in ways that are seemingly inconsistent, letting personal liberty trump objective mores in some cases, but objective mores trump personal liberty in others, and that these inconsistencies are expected and unavoidable just as they are in collective decision making.

Carl B
Aug 30 2016 at 9:17pm

During the discussion of the triage issue, it was claimed that there was no property right to the medical care that could be donated, bargained, or auctioned. Actually, there is a unmentioned property right, and in fact a 4th party in the room, discussion of which might have been illuminating. The property right actually belongs to the doctor (or possibly the hospital). It is his (or her) skill, and his time and effort — not to mention his liability and his reputation — that is being argued over. Any patient may have the right to refuse treatment, but without him nobody gets surgery. Any decision among the injured parties will then make a claim on him, and in your example would then force him to perform. The current system generally reflects that it is his choice, though rules and norms restrict him to make a judgement about who can be helped the most.

Can the doctor be taken out of the equation? What if it is a machine that performs surgeries? Well, that machine still belongs to somebody, and the patients don’t necessary have the right to force a choice over its use.

Is there a way we can give the property rights to the 3 parties, and not in a way that makes a claim on a 4th party? What if each of the 3 is able to assist in an operation, and an operation requires 2 people to perform? (Instead of their contribution to the operation, these could be chits or some other representative of their property rights.) Then they can bargain! In the original example Leo Katz suggested that Al’s ultimatum (“if you don’t perform on Chloe then you have to do it on me!”) makes Bee’s claim irrelevant. In my modified example, Bee is quite relevant. Without her, no surgery can be performed. For example, if she wishes to take a more abstract view, she may be able influence Al to take the surgery — “if you don’t take it, I won’t give it to Chloe.” Of course in the prior example the doctor might also have a theoretical right to refuse unless the patients accept his choice.

Of course, economists have it all over philosophers, because they can nail down the equities by introducing some math — what if we give some weights to the costs, benefits, and probabilities associated with each choice?

Aug 31 2016 at 8:56am

It is hard to imagine there is a “preference” for third payer since at no time was this system intentionally “chosen”. It came about with lots of laws and lobbying most of which were done with no intention to create a system that matched peoples preferences but to match the preference of politicians, hospitals, and others who had other ends in mind. Its comforting to imagine a government imposed solution is the right one, but it is as equally unlikely as we all have a preference for having our taxes taken from our paychecks. The laws exist to serve others needs.

Following up on Carl B and some of what was touched on in the podcast, in real life triage situation the choice do you want treatment or no treatment, the doctor does not care what you think about the triage beyond that. So, if Al refuses, he is out of the picture and if there is only Carole she gets treatment because she is next, if there is Bea she gets treatment. There are also a few more interesting scenarios. What if Al and Carole talk and agree that Carole needs treatment first (we will assume Carole is as selfish as Al is noble). But Al passes out before the doctor gets there. In this situation the doctor would treat Al despite Carole’s insistence that Al said she should be treated first, and against Al’s own wishes. Al could also be treated first because the physician could conclude that he was incompetent from pain or pain medications and not rationale (there is alot more ethical leeway in emergency rooms than the rest of medicine). These situations move more into medical ethics but some of the laws supporting them arise from the sort of conflicts discussed in the podcast.

The discussion on elections was particularly interesting this year. Both parties nominated someone with little appeal outside their hardcore supporters. On the republican side it seems like a clear illustration of how having increased options changed the outcome. All the more evidence we should avoid collective choices when possible, and when we make them ensure they have as little impact on our lives as possible.

Nathanael Snow
Aug 31 2016 at 4:57pm

It occurs to me that in many of the situations explored in this discussion that Adam Smith’s impartial spectator would side with the perversities in the law identified by Katz rather than what a Max U efficiency optimizer would advocate.

The reason Al wants to gift his property in order of treatment to C is that there is a personal and partial attachment. Bee is not party to that faction, so she is left out. The doctor could side with Al and C, but in so doing would become a part of that faction.

We want doctors to be as impartial spectators in how they treat our ailments, but personal in how they relate to us, thus the confusion.

The law I think is right in requiring impersonal order of treatment.

Aug 31 2016 at 5:52pm

Thanks for another thought-provoking episode. Really enjoyed the time spent on the choice/order dilemma and the steak/chicken/fish example.

I know you haven’t wanted to delve into specifics when discussing politics on Econ Talk this year (e.g. Trump v. Hillary v. Johnson), but I do hope you have one guest who will speak to some specific aspect of the race as the election nears.

John Alcorn
Aug 31 2016 at 9:50pm

At the risk of being uncharitable, and of displaying my ignorance, I would like to push back.

1) Although the title of this episode of EconTalk (and of Prof. Katz’s book) is Why the Law is so Perverse, the examples that Prof. Katz adduces hardly seem “perverse.”

In the context of law, a definition of perverse is: “(of a verdict) against the weight of evidence or the direction of a judge on a point of law.” A general definition of perverse is: “contrary to the accepted or expected standard or practice.” Conventional medical triage, the prohibition of torture bargains in sentencing in criminal law, and the imposition of damages rather than specific performance in contract law, are not perverse by either definition.

Rather, the hypothetical alternatives, sketched by Prof. Katz, are perverse. Indeed, Prof. Katz himself states that the torture-bargain arrangement “would be perverse”!

2) Does the voting paradox (preference cycling) really have anything to do with the rules of medical triage, the ban on torture bargains in criminal law, and the imposition of damages rather than specific performance in contract law? All of these legal rules conform to robust moral intuitions of large super-majorities of voters.

Kenneth Arrow’s impossibility theorems don’t shed much light on these cases. As noted by commenters Michael and Nathanael Snow, Ronald Coase (transactions costs) and Adam Smith (the impartial spectator) are much more relevant.

I would add, by way of conjecture, that John Rawls’ veil of ignorance, which Michael Munger discussed last week, might produce legal rules akin to the Kidney Club. Note, however, that kidney allocation is not only a supply problem, but also a matching problem! Supply by lottery would have to be designed to square with the need for timely individual matches.

3) Re: Torture bargains in sentencing in criminal law, Prof. Katz writes: “So we’ve got this bargain that seems beneficial to the state or the law enforcement authorities and indeed to the public at large, because they get to achieve retribution and deterrence by much cheaper means than lengthy confinement.” Prof. Katz neglects another, crucial purpose of incarceration: incapacitation. Steven D. Levitt finds that incapacitation played a large role in the major reduction in crime in the USA in the 1990s. See here. In light of the incapacitation effect and repugnance, there is no puzzle why the law prescribes incarceration rather than much briefer, consensual torture of the offender.

Sep 1 2016 at 12:33am

I think Carl B is right on with pointing out the property right belongs to the doctor. It’s even more clear when they brought up the “even more extreme” case of Al electing to play chess with the doctor instead of treating anyone. The example neglects the fact that there are actually many doctors choosing to spend their time in other ways than being on call in this ER. At the time this example is occurring, I’m sure several doctors are enjoying a casual dinner with their spouse, spending time with their children, or even (god forbid) playing chess with each other.

The decision is the doctor’s and the doctor’s alone, if he is swayed by Al’s argument, we might find that objectionable, but that is not “our” call. If the doctor instead decided to not show up that day (or his shift ended right as all 3 entered the ER), it would be hard to argue that he is obligated to provide his labor services based solely on others need. If this were true, that slippery slope would end with doctors simply being slaves to “us”.

Perhaps this is missing the point of the exercise, because it is still worth considering what is the most moral decision for the doctor to make give the fabricated constraints. However, I still believe it’s a crucial point that the doctor has the additional option of doing nothing (and since there is more than one doctor on the planet, many must be exercising this option during the example). Whatever decision the doctor makes belongs to him. We may disagree with whatever decision he makes, but we cannot use force to change it. If we believe we do have the right to decide for him, then we are also saying that we have the right to force other doctors to give up their leisure to treat the other two patients. I consider this an unacceptable moral stance.

S Powers
Sep 1 2016 at 10:28am

I object a little bit to the verbal sophistry of starting a point about contracts for organ donation between equal and willing partners, and then moving onto indentured workers contracts.

Those contracts, moving people from the Old World to the New were hardly formulated without significant coercion, violence, and all manner of systemic inequities. To spend time discussing how those were enforced would seem to grant them a degree of validity that seems very odd to me given what we know about how empires functioned. I think most of us, specially the brown and black ones, are certainly grateful that the law has evolved since then.

Thank you though for making me sit up over my morning tea to have a moment of post-colonial rage.

Russ Roberts
Sep 1 2016 at 3:20pm

S Powers,

Sorry to ruin your morning tea.

I would make a strong distinction between indentured servitude and slavery. From the Wikipedia entry on “Indentured Servants”:

Until the late 18th century, indentured servitude was very common in the British colonies of North America. It was often a way for poor people in Europe to emigrate to the American colonies: they signed an indenture in return for a passage. After their indenture expired, the immigrants were free to work for themselves or another employer.

In some cases, the indenture was made with a ship’s master, who on-sold the indenture to an employer in the colonies. Most indentured servants worked as farm laborers or domestic servants, although some were apprenticed to craftsmen.

The terms of an indenture were not always enforced by American courts, although runaways were usually sought out and returned to their employer.

About half of the European immigrants to the American colonies in the 17th and 18th centuries were indentured. Between one-half and two-thirds of white immigrants to the American colonies between the 1630s and American Revolution had come under indentures.[1] However, while almost half the European immigrants to the Thirteen Colonies were indentured servants, at any one time they were outnumbered by workers who had never been indentured, or whose indenture had expired. Free wage labor was the more prevalent for Europeans in the colonies.[2]

It wasn’t without problems. But outside of kidnapping, indentured servants had very different lives from slaves:

Not all European servants were sent willingly. Several instances of kidnapping for transportation to the Americas are recorded. An illustrative example is that of Peter Williamson (1730–1799). As historian Richard Hofstadter pointed out, “Although efforts were made to regulate or check their activities, and they diminished in importance in the eighteenth century, it remains true that a certain small part of the white colonial population of America was brought by force, and a much larger portion came in response to deceit and misrepresentation on the part of the spirits [recruiting agents].”[6] One “spirit” named William Thiene was known to have spirited away[7] 840 people from Britain to the colonies in a single year.[8] Historian Lerone Bennett, Jr. notes that “Masters given to flogging often did not care whether their victims were black or white.”[9]

Indentured servants could not marry without the permission of their master, were subject to physical punishment (like many young ordinary servants), and saw their obligation to labor enforced by the courts. To ensure uninterrupted work by the female servants, the law lengthened the term of their indenture if they became pregnant. But unlike slaves, servants were guaranteed to be eventually released from bondage. At the end of their term they received a payment known as “freedom dues” and became free members of society.[10] One could buy and sell indentured servants’ contracts, and the right to their labor would change hands, but not the person as a piece of property.

The American Revolution severely limited immigration to the United States. But economic historians differ on the long-term impact of the Revolution. Sharon Salinger argues that the economic crisis which followed the war made long-term labor contracts unattractive. His analysis of Philadelphia’s population shows how the percentage of bound citizens fell from 17% to 6.4% over the course of the war.[11] William Miller posits a more moderate theory, stating that “the Revolution (…) wrought disturbances upon white servitude. But these were temporary rather than lasting”.[12] David Galenson supports this theory by proposing that the numbers of British indentured servants never recovered, and that Europeans from other nationalities replaced them.[13]

Craig S.
Sep 3 2016 at 6:02pm

Hi Russ,

Very interesting discussion with Leo Katz. I enjoyed it very much.

However, I have a quibble. Regarding the example provided by Mr. Katz regarding the triage of Al, Bea, and Chloe, I would take issue with one of the premises. It seems to me that the decision as to which of the patients to treat is an allocation/fairness question which would be the responsibility of the physician on duty. I don’t see it as a “right” possessed by Al, and thus transferable. He didn’t enter the emergency room with a right to first priority treatment any more than a potential kidney transplant recipient would have the right to transfer their position on the priority list to someone else (except by declining to accept a kidney). His priority arises from an externality, his medical condition at the time he arrived, based on rules established (formally or informally) by the doctor or the institution. Thus, I don’t see how that could endow Al with a right to transfer that right to his wife or anyone else.

This doesn’t take away from the interesting points that Mr. Katz makes, but I do think it undermines the pertinence of this particular example.

Sep 5 2016 at 9:26pm

While there was interesting potential in the topic, it seems to me that some of the illustrations did not serve well due to problems that made them not relevant to the point.

For example, even before reading the other comments, I had already come to the same conclusion as Carl B, Kevin, Adam, and Craig S. “Al” has no rights in the triage matter that belong to him in any way that he can transfer as he pleases.

Leo Katz actually exposed the problem with this concession:

“For that matter, if Al just refuses treatment, well that just leaves Chloe anyway.”

Exactly so. Al has the right to refuse treatment, in which case treatment goes to the next person in triage priority according to the policies of the doctor/hospital providing the service. Al never had a right to transfer treatment priority according to his own preferences. So the illustration fails and the point that might have been worth making is not served by it.

Another example that I believe fell short was the illustration of one man wanting to punch another. The second sets up terms that the first does not accept. But then the first punches the second anyway.

Can the second expect the “contract” or “agreement” to be enforced? No, of course not. There never was agreement. They never came to terms that they both accepted. Therefore, this becomes nothing more than a case of unilateral action by the first person. He could be charged with assault and battery, but not for breaking an agreement. As far as I can see, the example fails as an illustration of any of the points that I believe Katz wanted to promote.

Finally, there was an attempt to distinguish the unenforceable agreement illustrations from other cases where your body is being threatened (which justifies a certain level of lawful force). I could not help being distracted by the fact that if one needs a kidney to live, failure to honor the kidney agreement does indeed threaten your body with death. However one might try to rescue that contrast, this path for illustration doesn’t seem to have the degree of clear distinction I suspect was intended by the line of reasoning offered.

For myself, I would have liked to hear more about actual examples of cycles of preference and priority in law.

Even so, thanks for yet another thought provoking talk.

Sep 6 2016 at 7:54pm

Leo Katz made the following claim regarding voting, which I believe turns out to be incorrect or at least misleadingly incomplete.

But 200 years later someone investigated the question: Can we construct a system that doesn’t have these manipulation perversities, that doesn’t have the dependence on irrelevant alternatives, that doesn’t have cycling? And, you know, the famous economist Kenneth Arrow made a startling demonstration that that’s impossible. It’s going to be either cycling, or dependence on irrelevant alternatives, basically. He made some very simple background assumptions that are not worth going into, basically that this is a truly collective decision as opposed to being a dictatorship. But once we make those very elementary background assumptions, it turns out that one of these perversities, either cycling or dependence on irrelevant alternatives, is going to be present; and that therefore this possibility of manipulation is also going to be ubiquitously present.

Arrow’s assumptions do matter because they restrict consideration to “ordinal” voting, i.e. where people rank the candidates by 1st choice, 2nd choice, etc.

Arrow’s result does not apply to “cardinal” voting methods, such as range voting (a.k.a. score voting) where each voter scores or rates each candidate.

Range/score voting provides more information than ordinal voting. It succeeds regarding independence of irrelevant alternatives and has many other desirable qualities.

In summary, range voting satisfies the monotonicity criterion, the participation criterion, the consistency criterion, independence of irrelevant alternatives, resolvability criterion, and reversal symmetry.

As it satisfies the criteria of a deterministic voting system, with non-imposition, non-dictatorship, monotonicity, and independence of irrelevant alternatives, it may appear that it violates Arrow’s impossibility theorem. The reason that range voting is not regarded as a counter-example to Arrow’s theorem is that it is a cardinal voting system, while the “universality” criterion of Arrow’s theorem effectively restricts that result to ordinal voting systems.
Range Voting Properties

Range voting also minimizes Bayesian Regret, which some would say is the “gold standard” for comparing single-winner election methods.

Please see also:

Sep 8 2016 at 10:16am

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.–Econlib Ed.]

Attila B
Sep 10 2016 at 2:14am

Thank you for this great podcast. I’m a long time listener (Econtalk is the highlight of my morning stroll to work) and first time commenter.

Don’t have anything nearly as intelligent to say as the other commenters, but just wanted to draw you attention to this: https://youarenotsosmart.com/2016/07/18/yanss-080-deep-canvassing/

It’s about “deep canvassing” – a way to change minds about quite sensitive issues (like LGBT marriage). The academic work is not without problems (as is covered in the linked podcasts) but work is ongoing and seems promising.

Live long and prosper!

Robert Swan
Sep 10 2016 at 7:57pm

Most of this discussion played on my own leitmotif: that when making comparisons you must transform all the multi-dimensional complexities of the things being compared into one dimension, an axis running from worst to best.

In the Al, Bea, Chloe situations, one axis was used to make the decision while a different axis was used to conclude that that decision was “perverse”. Hardly a surprise.

It seemed that Prof. Katz understood this just fine too. His example of a buyer deciding between cars was explicitly to map multiple criteria down to a simple ranking.

Later on though, he and Russ made me doubt how well they understood the core problem. Both suggested that a dictatorship is immune — that the problem all hinges on this “collective choice” business.

Choice, collective or not, is messy. The dictator still has to distill complex alternatives down to simple priorities. He is spared the next step of reconciling his priorities with others’, but he is in much the same position as Katz’s car buyer, and just as prone to buyer’s remorse.

With that off my chest, the talk was well up to standard; a good laugh at the “in that case I’ll have the chicken”, and the comments have been well worth reading. I like the new transcript layout too.

Sep 14 2016 at 11:32am

An interesting podcast covering some salient topics.

I just came by to link to what I think is the best article on self-defense and distinguishing it from court room tort case.


Comments are closed.


EconTalk Extra, conversation starters for this podcast episode:

This week's guest:

This week's focus:

Additional ideas and people mentioned in this podcast episode:

A few more readings and background resources:

A few more EconTalk podcast episodes:



Podcast Episode Highlights
0:33Intro. [Recording date: July 2, 2016.]

Russ Roberts: Now this is an utterly fascinating and challenging book. It's full of logic puzzles, incredible insights about law, collective choice. And I want to warn parents who may be listening with young children, we may get into some adult topics. We're going to start with some examples of what you mean by perverse--why the law is perverse. And I want to start to start with the kidney club. How would a kidney club work?

Leo Katz: Yeah. Some time ago a philosopher had the idea: there's a shortage of kidneys, and you know, people are put on waiting lists and they sometimes don't get a kidney in time. So then he thought, well, what if a whole bunch of people were worried, presently don't have any kidney problems, but worried that in the future they might have. And they would like to make sure there's somebody available who might donate a kidney to them. Why not form something called a Kidney Club where everybody agrees that if somebody falls ill and needs a kidney, a lottery will be held and somebody will by that lottery be designated as the donor? Now, of course, we each have two kidneys: that's why we're able to give up one of our kidneys; and it apparently, so I've read, has relatively little deleterious effect on the health of the donor. In particular one might think, well, one kidneys don't function as well as two. Apparently that's not the case.

Russ Roberts: It's the risk of the operation. It's not a pleasant experience; there's a risk of complications.

Leo Katz: But for instance you might think, what if one kidney goes bad: don't you want to have a reserve as a second? And it turns out that if you get kidney disease it usually affects both kidneys, so it's not even as though, giving up that spare kidney you are giving up too much. So it's not--

Russ Roberts: I'm in the Kidney Club, so if I lose that kidney I can get another one.

Leo Katz: So you join this Kidney Club and you have as a downside you might be designated to be a donor. But it's not a huge downside--

Russ Roberts: Low probability.

Leo Katz: though a significant one. But on the plus side, if you need a kidney and you would die if you don't get a donation, or you'd be limited to dialysis, which has significant downsides, you are protected against that because there's a donor. So it seems like it's eminently desirable, or lots of people might want to join such a kidney club, because it's a bargain that everybody is going to benefit from. They give up a little bit--this risk of having to be a donor. And in return they receive this tremendous insurance that if they ever need a kidney, they'll get one. The puzzle with that is: Although people might want to enter into this Kidney Club, the law almost certainly wouldn't enforce such an arrangement. And even as a matter of morality, I think we actually feel a bit askance--we feel uneasy about it. We feel uneasy about people entering into the bargain, and even more so: if we now imagine the lottery is held, somebody is designated at this point, balks and he wants to back out.

Russ Roberts: 'Yeah, I don't like the club any more,' says that person. You see the temptation. 'I want to quit. I want to quit the club.'

Leo Katz: Well, actually these bargains are not going to work if someone suddenly can back out of it when he's the one who is designated. So the only way in which such a club is going to work is if we all kind of lock ourselves in. And, you know, make a commitment that if somebody at the last minute tries to back out, well, he's forced to comply. Maybe he's held down and--

Russ Roberts: anaesthetized--

Leo Katz: anaesthetized and one kidney extracted. Indeed when everybody is told about this in advance, 'Think about it: this is not going to work unless we all make this iron-clad commitment, even a commitment to be held down,' everybody sees the point of that and is willing to do that because, although that's an unfortunate thing to have to undergo, it's counterbalanced by the benefit they foresee from the insurance they received to get a kidney.

Russ Roberts: And you agreed in advance voluntary. As an economist my temptation is to say: You thought this was a good deal; we understand that every realization in life may turn out to be not such a good deal, but you, in sane mind, made the decision to be in the club, and when the outcome turned against you, you can't cheat. The law should enforce that deal. But the law, you are saying, does not enforce that.

Leo Katz: Well, and what's most striking: there are sort of some standard situations where both law and morality agree that bargains shouldn't be enforced: when the parties aren't competent or some don't know what they are doing or there's maybe a problem with bargaining power.

Russ Roberts: Deception.

Leo Katz: Or there's duress, or deception. None of that is present here. These are people who are of perfectly sound mind entering into a bargain that we completely understand rational people would want to enter into. It's not even the case, as with some bargains, where we think, 'Well, almost anybody who wants to enter into this, he can't be of sound mind even if they looked like that, because that's such a crazy bargain.' But this doesn't seem like that. It seems like a really rational arrangement that rational people might want to enter into. And yet, I think we all intuitively balk at enforcing it. But more importantly, it's pretty clear that the law wouldn't.

Russ Roberts: And you give the general example of these kind of contracts.

Leo Katz: The general rule actually about the law--the situation doesn't even have to be as extreme as the kidney contract. But in any kind of contract, when you commit to doing something and you later on try to back out, the law will not, as they put it, 'specifically enforce your promise.' They'll force you to pay damages for not honoring the contract. But they won't actually have the sheriff at gunpoint as it were enforce it or allow other parties to the contract to enforce it. There was--once upon a time, some specific performance was allowed--the famous indentured servant arrangements that were entered into in the 19th century or the 18th century, were people who were brought as domestic servants from the Old World to the New. Those were actually contracts that could be specifically enforced so that if the domestic service who was brought from England to the United States then fled, he could be forced to return to his position and serve out his term. And it was actually something that everybody understood was necessary for these contracts to work, because the domestic servant wouldn't have any money to pay damages. So the only way in which anybody would be willing to enter into a contract to bring a domestic servant to the New World would be if he got this kind of commitment. And yet they came to be regarded as repugnant: We can't have indentured servitude now. We can't have any kind of commitment that would involve specifically enforcing a contract, as opposed to trying to extract damages from a person.


Russ Roberts: So, you give the example of someone contracts to have their portrait painted. And if the artist says, 'I don't like you,' or 'I changed my mind,' or 'I'm too busy now' and refuses to do the portrait--

Leo Katz: Or 'I like the portrait so much I want to keep it. I don't want to give it over to you.'

Russ Roberts: Correct.

Leo Katz: Yeah. He's free to do that.

Russ Roberts: And, having read your book, I tell this to people, they say, 'Well, of course they can't enforce that. Because you could do a bad job.' But the court could enforce that, too. The court could say, 'You did not put good effort in. This doesn't seem consistent with your other paintings. You clearly did not fulfill the contract. Do it again.'

Leo Katz: There's that. But then there's another difficulty that's maybe more fundamental: it's one thing to say, 'Well, it's pointless to try to specifically enforce it because it's too hard to ensure that somebody does a good job.' But think about the person who wants it enforced. He understands--the domestic servant may be surly and the portrait may be bad, but for whatever reason he'd just as soon have half a loaf than none. So, he's all right with it. So, it's true we might say, 'We don't force you to accept specific performance because it's an inferior thing and you should be entitled to get money instead.' But if you're the beneficiary and you say, 'Actually I prefer specific performance under these circumstances: I don't view it as, as good as money even though it's going to be an inferior performance and maybe the court can't make sure that it's up to snuff; but I'm okay with that,' then that argument of course is irrelevant, and the mystery is: Why can't he insist on it? Especially if the other party in advance was willing to sign up for that? Another version, as it were, of the Kidney Club problem. In some ways I view it as even more perplexing, because the appeal of the bargain is so strong and yet our aversion to enforcing it is even stronger. Even stronger than in the kidney case is the example with which I open the book. I imagine somebody proposing a dramatically newfangled way of reducing the costs of imprisonment and the cost of punishment by re-introducing what he calls 'voluntary torture.' He makes it possible for anybody confined for long term to submit, if he wants to, in the alternative to a much shorter term of confinement, or maybe just a very brief one, during which he will have something extremely painful--some extremely painful treatment administered to him. In other words, some form of torture. And we'll make it ever so marginally more attractive than serving out the 30-year sentence, marginally, so that he is induced to engage in it; but we don't get much of a reduction in deterrence and retribution, and whatever other objectives we are pursuing with the punishment, presumably we could achieve through torture as well. So we've got this bargain that seems beneficial to the state or the law enforcement authorities and indeed to the public at large, because they get to achieve retribution and deterrence by much cheaper means than lengthy confinement. It's actually attractive, marginally attractive, to the prisoner, too; and of course if he doesn't want to, he doesn't have to do it.

Russ Roberts: He can stay in jail.

Leo Katz: He can stay in jail. It's only if he requests it; and if we make absolutely certain that he's totally competent to make that decision, only then will we let him. But of course we wouldn't dream of putting in place such a system. Now we don't even have this problem of he gave his consent at one time and he regrets it and later on he regrets it and will we let him back out, the way it happened in the Kidney Club. Now, it's, you know: When we administer the treatment, we only administer if at that very moment he is happy to have that happen in lieu of having to serve his 30-year sentence. And we would balk at such an arrangement. It would be perverse.

Russ Roberts: You might. Some of us might not. Some of us might not balk at it. But the law balks at it. The law says: This is not--we don't do this. It's not a legislative solution, say, in the United States. And it is similar to the first case, say, as an economist looking at it, because a voluntarily agreed on, mutually beneficial exchange--which is the standard phrase we would use--it makes both parties better off, and in general it makes both parties better off without deception, as you said, or without coercion or duress. It's voluntary. Why would we not allow such things?

Leo Katz: And it's worth underlining that it's in some sense even more mysterious than the Kidney Club. Because it's not advance consent. It's consent at the very moment--we will only say we will only start the treatment right now if you are okay. This is not somebody who later on regrets and tries to back out. It's somebody who at this very moment says, 'Please do it.' And yet we say we won't.

Russ Roberts: And even if we said, 'Halfway through, you can change your mind: we'll put you back in jail. We'll stop.'

Leo Katz: Right.

Russ Roberts: We don't allow that, either.

Leo Katz: Right.


Russ Roberts: So, that seems perverse in that both sides would prefer that outcome and it doesn't happen. The same is true of the Kidney Club. Talk about the triage cycle, which is even more dramatic.

Leo Katz: Now, before I do that, I just want to briefly allude to this point you made. You say, 'Some of us wouldn't.' Many people who are, would describe themselves as very libertarian in their outlook on freedom of contract and are inclined to enforce lots of agreement where other libertarian-minded people would strongly hesitate--even they are led to hesitate when they encounter the voluntary torture example. Not everyone. But many. So, to my mind that kind of underlines the puzzle that lurks there. That even people who feel strongly committed to freedom of contract on some of these examples, especially this one, without quite being able to explain why, think they are not quite willing to live up to their abstract commitment.

Russ Roberts: And there's something--even I find I'm somewhat uncomfortable about it. I think part of the reason is we worry that there could be at lack of judgment there, perhaps. Someone under that unattractive choice. It's related to Mike Munger's concept of euvoluntary. It's a person who is clearly under duress no matter what: even if they enter into it freely, it's such an unattractive option, to call it a free choice seems somehow perverse in itself. And we're saying this, by the way, without even the possibility of televising the torture. Which could both increase the deterrent effect, and add entertainment for sadistic and cruel people. I say that with tongue in cheek--perverse tongue in my perverse cheek. Okay. Let's move on--

Leo Katz: But you know, on this point, that maybe there's some sort of inherent duress in the situation. One thing that makes me strongly hesitate to think of it that way is that, you know, it's not particularly different, functionally speaking, from medical treatment. Here we've got someone who as it were is paralyzed or limited in his mobility, or at least the situation is not very different from somebody who is limited in his mobility and was offered a very painful medical treatment that will solve that problem, that will treat it. We wouldn't describe that as a case of inherent duress. And this is functionally very, very similar. Doesn't dissolve that intuition altogether.

Russ Roberts: That's a very clever argument. I'm not sure--it may prove too much. You may explain why we're so comfortable with third-party medical payment, that I think has some perverse effects on its own--

Leo Katz: That's interesting--

Russ Roberts: that stops us from having to endure watching people under duress make monetary decisions.

Leo Katz: True, true. There's that further wrinkle.


Russ Roberts: So let's--

Leo Katz: Well, all right. Let me turn to this triage example you've alluded to--

Russ Roberts: It's spectacular.

Leo Katz: Well, it's a highly artificial scenario I construct because despite the fact that it's so super-artificial and unrealistic, like many unrealistic examples, I think it has the virtue of making more salient something that the more realistic examples tend to hide. What I imagine is a case of triage in an emergency room. Initially I imagine two people arrive. They both have just been involved in a car accident. One of them I call Al; the other one I call Chloe. They've had this car accident, and Al has been severely injured, and the way he's been injured is he needs immediate treatment or else he's going to lose the use of both of his legs. And then there's Chloe, who has been injured much less seriously; in fact the injury is one that would affect the use of her hands, mostly the use of her little finger, which actually will be a big deal to her because she's a passionate hobby pianist and any reduction in her manual dexterity is greatly going to affect her enjoyment of her own piano playing. If I told you nothing more and question arose: Who is going to get treated first? Is Al going to be treated first? We'll assume that whoever is going to be treated second is going to suffer some consequences from the accident. Al might in fact if he's not treated first but second might lose the use of his two legs; Chloe for that matter might suffer significant reduction of her manual dexterity. So, it's a big deal who gets treated first; and if I told you nothing more, then it's kind of obvious that both law and morality would dictate that the only doctor who happens to be on duty would immediately devote all of his attention to Al, the more seriously injured. Now I'm going to add a wrinkle. I'm going to assume that Al and Chloe are a couple: they are married; Al is incredibly devoted to Chloe, is incredibly concerned about how unhappy she would be if her piano playing ability suffered: is so concerned that he absolutely insists the doctor treat Chloe first, even at the cost that will entail to him, Al. Needless to say, the doctor is going to hesitate initially and not want to do this. But if Al is absolutely insistent and if it's clear that Al and Chloe, they both want this, then he's probably going to do it, and probably think that he ought to do it. For that matter, if Al just refuses treatment, well that just leaves Chloe anyway. So that's where we're going to end up: he's going to treat the less serious injury first under these circumstances. Now let's add one further wrinkle. Let me introduce a third character named Bea. Bea has also suffered an injury, and her injury happens to be midway between the most serious injury of Al and the least serious injury of Chloe: let's say it's the sort of injury where, if she's not treated she loses--we'll be super-schematic about it--she loses the use of one of her legs. Now, what's the doctor to do? Because as soon as he's about to treat Chloe, Bea pipes up and says, 'I've got a much more serious injury: I lose the use of my leg. Surely you're not going to treat Chloe with her trivial finger injury before you treat me with my serious leg injury?' And the doctor says, 'You've got a point; so, yes, I'll redirect my efforts from Chloe to you.' At which point Al says, 'Hey, but I've got a much more serious injury than Bea so you've got to treat me first. I've got this two-leg injury.' At which point the doctor says, 'You've got a point. That's what I'm going to do.' So, as he's about to treat Al, Al says, 'But you know, I understand I've got treatment priority, but frankly I would prefer you devoted your attention to Chloe. That's what I would prefer, and that's what Chloe would prefer.' At which point Bea pipes up, 'I don't want that. I can't live with that. If she's going to get treated, Chloe, then surely I have priority over her.' At which point Al says, 'Wait a minute. Wait a minute. You're not going to get treated no matter what. It's either me, or it's Chloe. So, you're kind of out of it. And as between me and Chloe, the only relevant parties here, we both prefer that Chloe get treated. So that's what should happen,' insists Al. Well, is that what should happen? Because even if, having said all this, the doctor decides to treat Chloe, my guess is we still think, 'Gee, can she really be treated ahead of Bea who has this more serious injury?' So we are kind of in a funny cycle. Whichever patient the doctor chooses to treat there seems to be another patient who we can argue has a superior claim. Seems like Bea has a superior claim to Chloe because of her more serious injury. Seems like Al has the more serious claim to being treated because his injury is more serious than Bea's. And for that matter, Chloe it seems has a superior claim to treatment because both Al and Chloe want it that way: and it seems like that's the kind of contract that we usually respect: It benefits both parties; it doesn't injure a third party. We'll assume they both know what they are doing, that neither of them is incompetent, even if Al seems a little bit maybe too besotted with Chloe. And yet there seems to be a genuine problem. It's not that we would rule the possibility of going through with Al's wishes absurd or totally out of bounds. But the important thing to realize is here we actually have concerns here about a bargain, a mutually beneficial bargain between two people that has no bad effects on third parties and that is concluded between seemingly competent parties. So at least in this particular context we have a compelling situation in which freedom of contract doesn't seem to prevail.

Russ Roberts: Well, it's--there is an aspect of this stylized example which needs to be mentioned. Which is: there's no property rights here for the parties. That is, we don't--and this is part of what you are saying, of course--is that Al, by virtue of having the worst illness, the worst condition, he doesn't get dibs. He doesn't get property rights that he could sell or give away. And in particular there are no prices in general. Period. So, we don't say: Well, whoever pays the most gets the medical treatment. We do in many markets. But in medical markets, in a triage situation in the ER (Emergency Room), or, to take a weird example that would seem to be unrelated but is not: we let people, now--we didn't before--we let people give up their place on an airplane to people more desperate to fly when the plane is overbooked. We essentially sell that--not in the way we might normally do it. But, the example--some of these examples that come up are about places where property rights are either ill-defined or are explicitly ruled out. And that's what this case is, to this economist's ear.


Russ Roberts: But what's clever about what you do, and this is what I want to turn to next, which I think is just brilliant, is that this discomforting example, this unpleasant example, has an analogue to voting cycling and problems in public and collective choice. So, lay out the case for--explain what the voting paradox is. We've talked about it before on EconTalk, but explain it again for listeners who have forgotten or didn't hear it. And we'll apply it to the triage cycle.

Leo Katz: Yeah. Maybe, if I could just back up a little on this issue of property rights. Which, it's unclear what role it plays in this example. Because you could--even if you tried to inject property rights into the example--that is, you initially try to make it less of a dilemma by thinking: Well, Al has a priority, and that priority is like a property right; and if he wants to alienate that or give it away to Chloe and have her possess it, that's kind of, you know, that's what you can do with property rights. Why shouldn't you be able to do it? And when one phrases it that way, one initially thinks: 'Well, that kind of resolves the case. Now it no longer looks so unjust for Chloe to be treated ahead of Bea because she has this property right that she inherited from Al. Now, what makes the property right argument more difficult is that in general the law takes the position that when you are in the case of a hard choice--when some person is about to suffer some significant detriment--you know, there's a hiker starving in the woods and the only way he can survive is by breaking into a cabin and helping himself to its, you know, to the foodstuffs there, he is free to do that because life is more important than property. So, similarly here, one might say, well, yeah, she's got this property right to treatment, but yeah, property rights can be invaded[?] for more important reasons, and Bea is a position to invade it for more important reasons. So her argument against Chloe would seem to survive, even characterizing this as a property right. So it's not clear that the dilemma can be resolved by simply classifying it as a property right. Although it's--maybe.

Russ Roberts: I take the point in light of the law. I see that point. And of course the law, though, might require the backpacker or hiker to compensate the person.

Leo Katz: True, true.

Russ Roberts: So, to me it's a little like--we may get to it later, but you talk about loopholes, in the book. Obviously, there's a law against speeding; if you have a medical emergency and if you race to the hospital above the speed limit you'll eventually get an escort if you are lucky. Not a ticket. So, but carry on.

Leo Katz: Yeah. Yeah. There is this--we, the usual context in which economists most think about cycling is a rather different one. Although--which is the problem of voting. And the most notorious illustration of this sort of cycling phenomenon which we've encountered in this triage example, the most notorious example that economists think about, and political scientists who think about voting, arises in the context of voting. And the voting paradox, I mean, that's its most clear-cut illustration and that's how people's attention was first drawn to this phenomenon. And it's the following kind of situation. We've got three candidates running for a certain position. And the electorate is divided. And they have preferences between the three candidates. And each of them has a very distinct set of preferences with regard to the three candidates. One group wants Candidate Al, ahead of Candidate Bea, ahead of Candidate Chloe. Stay within those, keep those characters alive. And then there's a second constituency, and they most would like Bea, and their second choice would be Chloe, and their third choice would be Al. And then there's a third group: they would most like Chloe, and if they can't have Chloe, they would most like Al, and if they can't have Al, they'd like Bea. So, everybody--and let's assume that the constituency is, sort of one third of all voters have the first set of preferences; one set of voters have the second set of preferences; one set of voters have the third set of preferences--in other words everybody has first choice with one third of the electorate, second choice with another third of the electorate, third choice with another third of the electorate. Now, what's puzzling about it is that if we were to run these people against each other in a simple--not having all three compete with each other but have two of them run against each other--then we'd find that Al is preferred by 2/3rds of all voters over Bea and Bea is preferred by 2/3rds of all voters over Chloe. And one might therefore say, 'Well, a 2/3rds majority prefers Al to Bea and a 2/3rds majority prefers Bea to Chloe, obviously society likes Al better than Chloe for this position.' But if then were then going to check and run Al and Chloe against each other they'd find that actually Chloe is preferred by 2/3rds of all voters over Al. So we get this odd cycle here, whereby each candidate loses out to some other candidate by 2/3rds--not a trivial margin.

Russ Roberts: And it's not, just to emphasize, this is not a specific example. It's more general than this, unfortunately: that majority rule in pairwise comparisons does not lead to a single outcome.

Leo Katz: Exactly.

Russ Roberts: And this is appalling and just horrifying. When most people see it, they actually, I want to say they literally don't believe it.

Leo Katz: Yeah. They don't believe it.


Russ Roberts: We have such reverence for majority rule. That, obviously that's such a good way to make decisions. What could be more fair? What could be better? And yet majority rule--and I never thought about this till I read your book--it's Rock, Paper, Scissors.

Leo Katz: Yeah. Exactly that.

Russ Roberts: It's: Rock beats scissors; scissors beats paper; but rock doesn't beat paper: paper beats rock. And this nontransitivity is very unattractive. And a lot of political scientists obviously have explained some of the more perverse aspects of the political process as a response to this nontransitivity. Or, more importantly, a better way to say it is: Order matters. And so, if you don't restrict, you end up in an endless cycle without any reliable outcome, like your poor doctor in the treatment room. And so a lot of peculiar aspects of voting and institutions around the political process, it's been argued, are there to avoid this kind of endless spinning your wheels. But you apply it to the law. Which is very clever. So, carry on.

Leo Katz: Now, one might wonder, if there is cycling going on in both these contexts, is there anything that connects them other than the fact that we run into cycles in both cases? Is there some deeper reason why we should run into cycling with voting? And also run into cycling in this example which really doesn't have anything to do with voting: nobody is casting a vote. And the connection is kind of interesting, and I think illuminating. The reason we get cycling in voting has to do with the fact that we are combining the preferences of different voters. Starting with this example and expanding on it and following up on all sorts of questions raised by it, economists and political scientists were able to show that one runs into this cycling problem or some version of it with just about any voting system--any situation in which we try to combine the preferences of different voters. After thinking about the situation for a while or indeed rather quickly, people working on this recognized that this isn't just a problem about voting, but there are many other situations in which we do something analogous to voting even though there are not several voters involved. That is, even if we have a single person trying to make a decision, but what he's trying to do is trying to take into account a variety of different considerations. That often works as though we were trying to aggregate the preferences of a series of voters into one outcome. So, for instance, I'm trying to decide which of several cars to buy, and they have a variety of criteria which I focus on having to do with its look, having to do with its price, having to do with its safety. And from each point of view, different cars fare better than other cars. And so I rank them each with regard to each of these dimensions; and then I have to somehow combine each of these dimensions into a final ranking between the cars. Well, now that starts to sound a lot like trying to rank candidates who are running for a certain position in light of the different preferences that different voters have for them. And it therefore stands to reason that we're going to run into some very similar sorts of problems--in particular cycling when we are making what people often call multicriterial decisions--decisions based on a multiplicity of criteria. And in a sense, that's what the law is doing very often; and in particular it's what's going on in this triage case: We're trying to take into account a set of different criteria, each of which would rank the patients along different dimensions. There's kind of the abstract need dimension reflected in our judgment about the severity of injury. And then there's another set of considerations having to do with their preferences, in particular the fact that one of them or two of them want one treatment not to follow the abstract severity of the injuries but their preferences about and their affection for each other. And when we are trying to aggregate these two different considerations, that's like trying to aggregate the preferences of different voters; and just as that can give rise to cycling, this can give rise to cycling, too. So, it's not an accident that we run into cycling in both contexts. It really is--the underlying mechanism that gives rise to it turns out to be the same.


Russ Roberts: It's about making a choice, a decision, when that decision is multidimensional--

Leo Katz: Exactly--

Russ Roberts: And that's--but the applying it to an individual is really remarkable. When you think about the irrelevant alternative problem--that people might change their preferences when a choice is offered that they are not going to take is weird. It's very unsettling.

Leo Katz: Should I maybe say something about the irrelevant alternative and how that relates to cycling?

Russ Roberts: Yeah. Sure. And then the killer amendment, which is the example you use in the book which has a similar flavor to it.

Leo Katz: Right. There's--cycling obviously is one kind of disturbing perversity, whether one encounters it in voting or one encounters it in law. Now, it's interesting to see what the initial reaction was that people had when they saw that majority voting could lead to cycling. Indeed, already in the 18th century when the voting paradox was first noticed by two French social scientists, Condorcet and Borda, one of them then concluded from this that majority voting should be abandoned. And he proposed an alternative method. And the alternative method was he suggested that every voter rank his candidates and that we then, just, you know, calculate the average rank for each candidate; and that should be the way we proceed. And he pointed out that if we did that we would never run into cycles. We might run into ties and then maybe have to use some random device or some other way of dealing with ties. So, for instance in the case of the voting paradox where everybody comes in--a third of the people prefer one candidate first, a third of the people another candidate first, and so on--if we applied his method to that we would find that the three candidates are tied. Which is actually kind of sensible because all three of them are in an equivalent position. But he thought that that would be a method we ought to use, and we wouldn't ever run into cycles, he gleefully pointed out.

Russ Roberts: Which, I should add--which, as weird as that is that they would be tied, there is something a lot more attractive about that, because if you think about the Al/Bea/Chloe election: If you are in charge of which order that pairwise comparisons take place, you can determine the winner. Which is much more disturbing--

Leo Katz: Right, exactly--

Russ Roberts: than saying, 'Well, we'll flip a coin for the three people.' Basically you are saying the person who has agenda control, says, 'Well, it doesn't matter. Let's vote on Bea and Chloe first and then the winner faces Al'. That will let Al win. Whereas if I did it a different way--I can pick the winner just by choosing the order.

Leo Katz: Right.

Russ Roberts: And that's--it's not just the fact that it's somewhat disturbing that the order matters. It's that it's even worse: the person in charge of determining the order determines the winner, if you think about it.

Leo Katz: Yeah. And that is actually quite right. And it's a very crucial thing to realize, that in some sense it's not just cycling per se that's disturbing. It's the underlying possibility that once you've got cycling, basically you can produce whichever outcome you want by just lining up the candidates in the appropriate order. And when we now go back to the fix that Borda had suggested, what he didn't realize, and what his rival, Condorcet, equally gleefully pointed out, is that he had avoided cycling, he had not avoided this possibility of manipulation. Because this method of doing average ranks has the following awkward property. If we have all three candidates present at the same time, they all will be tied. But suppose we remove one of the candidates, or it turns out he's dead; or for some other reason, actually, disqualified. And we re-run the election according to Borda's method about ranking the candidates. Well, then, suddenly now they are no longer tied, because now we just have two candidates left and one of them will get two thirds. It's basically like a majority election now: one of them has two thirds of the support and the other one has just one third of the support. So, by simply removing one candidate, two candidates who were in an equal position before are now in a very unequal position, and one of them prevails. Voting theorists who underline the absurdity of the situation say this is almost like somebody going into a restaurant and he's choosing between chicken and steak; and he chooses the steak; and then he's informed that there's also fish on the menu, and then he says, 'Well, now I'm going to have the chicken instead.' That seems absurd; and yet this Borda method seems like that, because the relative position of two candidates is strongly affected by whether a third candidate is present or not and can be affected even more dramatically than in my example. In my example they were tied when the third candidate is present, and when he's not present one of them prevails. But there are cases where it's, to begin with one of them is ahead and then a candidate is removed and then the one who is ahead now is behind. And this is sometimes called the 'violation of the independence of irrelevant alternatives'--or, it makes an irrelevant alternative determinate of what the relative ranking of two possibilities is. And of course once we've got that, we still have the possibility of manipulation--not manipulation in the way it occurred before, but just by throwing in an irrelevant candidate who is not going to win anyway, but just by making him part of the race, as it were, we can completely reverse the relative success of the only two relevant candidates in the race. So manipulation really cannot be removed. Or at least it can't be removed by this method. But 200 years later someone investigated the question: Can we construct a system that doesn't have these manipulation perversities, that doesn't have the dependence on irrelevant alternatives, that doesn't have cycling? And, you know, the famous economist Kenneth Arrow made a startling demonstration that that's impossible. It's going to be either cycling, or dependence on irrelevant alternatives, basically. He made some very simple background assumptions that are not worth going into, basically that this is a truly collective decision as opposed to being a dictatorship. But once we make those very elementary background assumptions, it turns out that one of these perversities, either cycling or dependence on irrelevant alternatives, is going to be present; and that therefore this possibility of manipulation is also going to be ubiquitously present.

Russ Roberts: And that's why after Kenneth Arrow published his book in 1957 on collective choice, that's why from 1957 onward most decisions were returned to the private sector, because collective choice obviously is a cesspool. No, I'm just kidding.


Russ Roberts: But what I want to turn to--it is a startling but depressing truth that Democrats, Republicans, people who love democracy, people who love republics, people who love everything other than dictatorships have to confront: that collective choice is a lot messier than we might have thought, given that majority rule often seems so straightforward. It's not.

Leo Katz: Right.

Russ Roberts: Now, let's turn to this question of: What's this have to do with the law? You've shown a parallel. Is the fact that the law rules out the kind of--so you are suggesting, going back to the triage, going back to the Emergency Room, is that Al will get treated. He will not be allowed to cede his rights to Chloe. Bea will not be allowed to override Chloe. It's just going to be Al. Done. Do you see that as analogous to what happens in certain voting regimes where certain opportunities for manipulation are not allowed? And does that apply to the other examples that you gave of the torture of the prisoner and the Kidney Club non-enforcement of the contract?

Leo Katz: I don't think it's particularly the aversion to manipulation, that--I think in a situation like triage, there will be a strong temptation--I don't know that it will necessarily always be carried out--but I think there will be a strong inclination to say to Al, 'Either you get treated, or Bea is going to get treated; but you can't just make this bargain with Chloe or this Pareto-optimal arrangement with Chloe whereby you alienate or allocate your right to treatment to her.' There will be a strong temptation not to allow him to do that. Now, one could--it's not that it's absurd to give way to that preference. But whenever a cycle like that lurks or presents itself at least very starkly, there's one of several plausible principles that will have to be rejected. And when the competing principles are the ones we've got in play in the triage case, one could well imagine the freedom of contract principle losing out. Doesn't have to, but it's--one could see how, you know, there would be significant pressure to break the cycle in that way. And I think that pressure exists in a lot of other examples that are not transparently like the triage example, but that have sufficient similarity with it. In particular, there's this argument that Bea is able to make in the triage case, which we can see essentially being made successfully in a lot of other different contexts. Basically what she's able to say to the doctor who is thinking about whom to treat, she's able to say, 'Look, I can see giving up my one leg for the sake of having the two legs of Al treated. That's a justifiable sacrifice. I can't see being asked to make that sacrifice for the sake of Chloe's finger, even if Al wants that to be the case, even if in a sense I'm not disadvantaged because one way or another I'm not going to get treated: either Al gets treated or Chloe gets treated. But let's make it more extreme: Suppose he wants to alienate his right to treatment for, you know, playing chess with the doctor instead during that time. Surely I can't be asked to make a significant sacrifice for something as trivial as that.'

Russ Roberts: Or pays the doctor a sufficient amount to--I have to concede, I said earlier about property rights; but the truth is, most people are very uncomfortable with the idea of: Get Al out of the picture; Chloe comes into the Emergency Room and says, 'I'll pay a million dollars to save my finger,' and Bea says, 'Well, I only have $20.' Most people would say, 'Well that shouldn't be relevant. We don't like the idea of property rights.' And my idea as a property-rights-loving economist, I'd say, 'Well, that's why there'd be more than one doctor.' A situation with more freedom to choose, and not these rules, etc. But the truth is, is that we are uncomfortable. The idea that Al would pay the doctor and not operate on any of them. Forget Al. I come in. I say, 'Doctor, I know you've got these three people. Here's $200,000. Play chess with me for an hour and we'll let these other people all go their own ways.' That also is repugnant.

Leo Katz: Well, let's return to this problem about specific performance. We--let's think about what we're really doing when we give someone the right to specifically enforce a bargain on the basis of advance consent. The uncontroversial context in which we are essentially able to use force, in which we are able to exert force on someone, is when he is trying to do something to our physical body. Or, maybe, to property to that is maybe directly in our possession. Under those circumstances we are able to use physical, self-help remedies. If it's our body, deadly force; if it's just our property, well we can use some force but not deadly force. In addition, it's one of those situations where the criminal law steps in, we can summon the authorities to provide protection; we can summon the authorities to provide punishment for any invasion of our body or of property in our possession. When we try to--when we ask for specific performance, we are really asking for the right to treat a promise the way we ordinarily are entitled to treat our bodies and the way we are entitled to treat property in our possession. We ask to be able to use the same self-help measures to implement it. We ask, again, to be able to call on the authorities, the way we do to be able to protect our body. We might even try to be able to punish criminally the noncompliance with such a promise. The question is why shouldn't we be able to do that? Well, in a sense we are now trying to arrange it by a voluntary arrangement. By a contract that we enter into in the context of the Kidney Club or some other context. We're trying to ask the authorities, like the doctor, to treat a lesser interest that we happen to care about a lot as though it were a greater interest. Just as Al is asking the doctor to treat Chloe's finger injury as though it was as serious and had as much priority as his two-leg injury, we are, when we are trying to protect promissory interests with self-help, we are asking society to treat that interest as being equivalently serious as a threat of a physical injury to our body and to our property. And, well, you know, just as Bea might object to that, the person who is being asked to treat this lesser interest as being as serious might similarly balk. Might similarly balk for being asked to make a sacrifice that is commensurate with a more serious interest, like the two-leg injury and like your body and like your property, being asked to make that sacrifice now for the sake of a much lesser interest--namely, a promissory interest. It would be a promissory interest that we value highly; but it frequently happens that something we value very highly, we in fact have a low interest in. It's this peculiar phenomenon--that we often have a low interest in something that we have a high preference for--that leads to this oddity, this inability to strike a bargain we want to strike.


Russ Roberts: So, I didn't quite follow that. So, give me--let's take either the Kidney Club, or the painter who has promised to paint the portrait--where is the conflict in the principles that's akin to the triage cycle principles? In the triage cycle, it's--there's two principles. There's the: Worse injuries should be treated before lesser injuries. And then there's: If you have a mutually beneficial exchange, we should allow it. So, in the case of the triage cycle we argue--we--the law--you argue, argues that the worst injury principle should dominate the mutually beneficial exchange between Al and Chloe; and therefore Al should be treated, and Bea is out of the picture. How does that apply--I don't see how that applies to the Kidney Club or the painter who wants to renege on his promise of painting [?].

Leo Katz: Well, the painter essentially is saying, or the two parties to the agreement essentially say, 'Well, you know, we would like the promissory interest that the other party has in my painting, the painting, or giving up whatever else it is that I've promised: we want that to be treated as though it were already property in the promisee's possession. We want that to be treated as though it were an invasion of the victim's body.' With the same level of seriousness, they ought to be able to demand from the rest of society, or indeed from anyone whatsoever the same level of respect for that interest that they ordinarily get for their body. Just as Al is able to say, 'Because I have treatment priority for my two legs, I would like to substitute for that Chloe's treatment of her finger. I want that to be viewed as being as serious as my interest in my two legs, because I happen to value that more than my two legs. I am entitled of my two legs. I value the treatment of her finger more than the treatment of her two legs. Therefore I would like her finger injury to be treated as seriously as my two-leg injury. In return I forego the treatment of my two legs.

Russ Roberts: So how does that work--let's have--I'm the dignitary and you are the artist. And I've hired you to paint my portrait. And I'm going to hang it in my estate in Kensington. Where is the analogy here? Help me out.

Leo Katz: Yeah. The painter is saying, 'You know, ordinarily I can't be--the only circumstance under which I can be subjected to significant physical force or intrusion by this other guy, the guy whose portrait I'm supposed to paint, is ordinarily if I do something to his body, or if I steal something from him. ' Those are the only circumstances ordinarily in which he's able to use physical force against me or to call in the sheriff to force me to stop or even to punish me. If I do other, more minor things to him, such as not keep an ordinary promise--an ordinary promise to sell something to him--he can't use physical force against me. He can call the authorities to force me to comply. But not [?] talking about a case in which he makes a specific promise, a promise specifically enforced--just an ordinary contract, the painter says, if I were to promise to sell him something and then I don't follow through, he can't say, 'Well, I've got a right to this stuff and since you are not giving it to me, I'm going to take it.' We say, 'Your right, the strength of your claim to whatever it is that's been promised to you, is not at the same level of the strength of your claim to your kidney or the strength of your claim even to property that happens to be in your possession. And therefore you can't use force to implement that right, to, you know, vindicate that claim. And you can't even ask the authorities to help you vindicate it, except you can ask for damages. But you can't do anything more. And you can't have me punished or thrown in jail for not honoring this promissory right that you've got.' Now, what you're trying to do is you are trying to arrange things in return for, or [?] the painter--try to arrange things by, I accept that money in return, I accept that money from you, and in return I try to make this promissory right you've got--we try by mutual agreement to make it such that the promissory right will be treated by society at large--and for that matter, by me--as though it were your kidney still inside your body. As though it were some property that you possessed. When Al is trying to alienate his property to Chloe, Chloe is, both he and Chloe are basically asking that something that a weaker claim as a result of an agreement that they've struck, be treated like a much stronger claim. Like a much stronger priority. And the same sort of arrangement, the painter and the client are trying to strike when they want to set it up in such a way that this promissory claim that the client has is treated like a much stronger claim, one that can be vindicated by, you know, by self-help. By something equivalent to self-defense. Even though it's a promissory right that ordinarily can't be treated in that way. And as in the triage case, you can't artificially as it were inflate the certain claim into a much stronger one by mutual arrangement. You can't do that here.


Russ Roberts: So, let me see if I understand. I'm still a little bit confused. So, I hire you to paint my portrait. You show up at my house. I sit down. You start painting, and after 15, 30 minutes you say, 'This just isn't working for me. I've got a lot on my mind. I really don't have time to paint this right now. I've got other things I care about more. I just have to back out of the deal.' And I go to the law, and I say, 'You didn't keep your promise; and therefore the law should use the threat of force or jail and force you to paint my picture.' So, that's one principle. The principle is: Contracts should be enforced. Is the second principle then that the enforcement of the contract has an immoral side to it because you are equating the painting of a portrait with the violation of property or person? Is that what you're claiming are the two principles that conflict here?

Leo Katz: Yeah. You've got--you have--let's see if--this may be another way to make the point. Suppose people wanted to arrange it such that we all have--for the moment--the right of self-defense. Which is very similar to the right to specifically enforce. Or at least it's that analogy that I'm playing off of. The right of self-defense is ordinarily limited in these various ways: you get to use a lot of force for your body, less force for property, and no force at all for lesser kinds of interests. But suppose people want to modify that by agreement: they want to set it up so that they allow more force to be used for lesser interests. Or, one person says--

Russ Roberts: How about this? I put on my front door: 'Warning! If you steal from me, I'm going to shoot you.' I'm going to shoot with deadly force. There's these classic cases in the law of a trip wire that's going to--

Leo Katz: Spring on you--

Russ Roberts: Spring a gun that's going to either shoot me in the legs or shoot me in the heart. Shoot you in the legs is one issue; but shooting in the heart, for somebody who is stealing a phone is not considered just. We'd all agree that's unjust--you shouldn't be killed for stealing a phone. But, I post that outside my house. Right? I say, 'If you come in here, your risk deadly force.' And the person enters voluntarily.

Leo Katz: Yeah. Let me--

Russ Roberts: Is that your point?

Leo Katz: It's a nice example. And there's a version of the example, and sometimes makes the point even more salient, that I sometimes use in class to make the same point. So, let's not make it breaking into a house. But it's often these things are clear when we actually use physical force on someone's body. So, suppose one person approaches another, and he says, 'I notice you are a body builder and I've been told that people who do that have these iron-clad stomachs, and even if they are hit hard in the stomach they can bear it and they don't flinch. And they don't get much injured. I'd like to try that out.' And the bodybuilder says, 'Okay. Strange request, but yeah, I'm all right with that, for a certain sum of money.' And then the other guy asks, 'Well, how much?' And he says, '$500,000.' And the other guy says, 'That's ridiculous. That's way high. I don't want to pay $500,000.' And the bodybuilder says, 'Suit yourself. You want to do it, $500,000. Otherwise, I don't give you permission.' At which point, the other guy takes aim; hits him in the stomach. The bodybuilder says, 'Okay. I guess you owe me $500,000.' He says, 'No I don't. Go ahead, sue me for the tort defending you. And you'll get some money. You won't get $500,000 for that.' And I ask: Well, shouldn't he? And people initially say, 'Well, yeah. Of course he should. I mean, he can name his price for being hit; and the guy chose to hit him. So of course he owes him $500,000.' And then I follow up with an example that's kind of like yours, and I imagine somebody driving around with a car and a sign on which says, 'You dent this car, you pay a million dollars.' And if somebody dents the car, intentionally or not, do we then think that he owes a million dollars? No. We think he owes whatever amount we think somebody is entitled to, as a matter of tort law, not as a matter of contract law, for the injury that he suffers. So, we've got this other scale, as it were. We've got the scale of tort law and of criminal law that rates, that values things at a certain level and that seems to determine how much we get by way of damages. And how much we get for an injury or how much we get for a trespass. Like the injury against the hitting of the body builder. And then there's this alternative set, the voluntary agreement, the contract set of principles that follow preferences. And it's the combination of these two--the ranking principles of tort law and of criminal law--under which we don't say you get a million dollars simply because your car is dented, just because you put up a sign. You can unilaterally modify the valuation scale: that determines how protected your various interests are; that determines how much you can suddenly decide, 'I value my car so much, I value it more than my body; therefore I get to use deadly force to defend it against anyone who threatens to damage it.' So, when it comes to criminal law and tort law, we've got this scale of valuation, very similar to the scale of valuation involved in the Emergency Room. And then of course we've got this other scale of valuation that we use when people strike bargains. And then there are these contexts where the contracts and the tort scales, as it were, are in competition with each other or need to be combined or we need to strike a balance between them. And when we do, one of them loses out. And sometimes it's the contract approach that loses out. As it does in the example of the bodybuilder and as it sometimes might do in the Emergency Room and might do in the case of specific performance. It [?] always lose out, I should point out--

Russ Roberts: So, I--

Leo Katz: Sorry. Go ahead.


Russ Roberts: No, sure. But one way to think about this, to keep it in the painter, see if I have this right, your point about the bodybuilder, which is really interesting. So, I hire you to paint my portrait. I give you a million dollars. And you show up and you start to paint the portrait and you realize you don't want to do it, or other things come along. And you say, 'I'm sorry, I'm not going to do it. Here's your million dollars back.' And I say, 'But that's not what I want. And that's not what we agreed to. We didn't agree that if you changed your mind, you'd give me back a million dollars. You agreed that you'd accept $1 million; and you'd paint my portrait. So please do so.' And so, if I take you to court: 1. The court will not let--under threat of torture--that you paint a great portrait of me. That's number one. And number 2, and that's the subtler thing I see now, number 2, my damages will not be a million dollars. You will not say, the law will not say, 'Obviously I valued this portrait at, at least a million, in fact more than a million or I wouldn't have entered into it, and therefore the damages are not going to be a million dollars for my foregone pleasure from the portrait not being executed.' So, it seems to me that you are saying that, this point that there are two principles, effectively, in what's been lost. And that the law is not going to always invoke the monetary value of the lost thing, whatever it is. Is that a good way to say it?

Leo Katz: Well, or at least the subjective monetary value. It will put monetary value on the matter.

Russ Roberts: Correct. Good point.

Leo Katz: But it won't say, 'Just be you happen to value it at, that's what you are going to get. The claims, it seems, are often, as it were, objectively evaluated. And, though that at first sounds paternalistic, it actually isn't, particularly. We can play around with a lot of contexts where it will become apparent that most of the time we make moral judgments, we cannot help--we can't completely go with someone's subjective valuation. I always find the law of self-defense particularly compelling in that regard, because we wouldn't dream of saying just because somebody values something much more than his body, therefore he gets to use deadly force to defend it. Instead we take the objective view according to which we say, 'It still is just a trinket, and you can't use deadly force to defend a trinket.' And as soon as we've got that--as soon as we therefore objectively value something, certain things, in a way that might therefore deviate from the subjective valuation--as soon as we say people have an interest in something that could be lower than their preference, or they have an interest in something that could be higher than their preference--as soon as interests and preferences get out of kilter, we've got a problem with bargains. And with respecting bargains.


Russ Roberts: So, we're out of time. Let's close with--try to summarize this. There's a lot of interesting thought-provoking material here, and the book is even richer. But it seems to me that the--what you are trying to argue here is that there are many things in law, or in social norms even, that are perverse. That seem like: Wait a minute. Why are we not allowing this thing to happen? And you are arguing that, even though there is a principle that suggests that it should happen, we often forget there's another principle that is in conflict. And that that principle sometimes gets precedence. And sometimes not. But often gets precedence. And that what looks to be a perversion is actually sensible. Is that a good way to summarize it?

Leo Katz: Absolutely.

Russ Roberts: Go ahead. Take your shot.

Leo Katz: Very good. Right. No, no. That puts it nicely.

Russ Roberts: Would you want to add anything else?

Leo Katz: No, I don't think I could improve on that. At least I gave it my best shot. But this, I think, puts it in a nutshell.

Russ Roberts: Well, the book made me think a lot. As did our conversation. And I think how often it is the case that we as economists, legal scholars, armchair moralists--we think that something is "obviously a good idea." And your book reminds us that sometimes it's a little more complicated.