 We now move to the fourth paper in our program, David Friedman, who has confided to me that he had a rush of political correctness and he wanted me to eliminate why not hang them all from his title and to substitute the civil criminal puzzle or should we abolish criminal law? David is Professor of Law and Economics at the University of Santa Clara School of Law. He was not at least, I think technically rather like Gordon Tulloch, he didn't have a lot of formal training in economics. It does seem to be a prerequisite for doing great work that you don't, maybe something wrong with our education system or you might say something right about David Friedman. David has of course published widely in academic journals in both economics and law. His undergraduate textbook Price Theory is a Best Seller as is his treatise on free market economics, The Machinery of Freedom. If you are interested and have a deep appreciation for economics, I strongly recommend that you read David's most recent book, A Hidden Order, The Economics of Everyday Life. In addition to these works, he's written on the economics of private law in Iceland and common law in England. He received his PhD in physics from the University of Chicago and then he snuck back into economics through the back door and into law and economics, of course, where he is a leading figure. So David will talk now and we'll follow the usual rules. Just paging to the top of my talk. I should say with regard to the titles, I had proposed two different subjects I was willing to talk on. One of them is the one I'm going to talk on and the other one, the title was Why Not Hang Them All, The Inefficiency of Efficient Punishment. And that I think is going to come out in the JPE as part of a special issue that's coming out in a year now. But I thought I would give you the civil criminal one, which is maybe more interesting and which has a few fragments of my article on Posner that Paul somehow missed when he was giving his talk. We start out with a very simple puzzle and that is that if somebody breaks your arm, you call a policeman, but if he breaks your contract, you call a lawyer. That is, we have what look like two parallel legal systems, one of them criminal, one of them civil. And that raises a series of interrelated puzzles. One of them is why I have two systems. Is there any good reason to? Could we do just as well by using, let us say, the civil system to cover all of the things now covered by civil and criminal or perhaps the other way around? And as some of you know, my article on Saga Period Iceland was partly motivated by the fact that that was the society in which, in effect, what we think of as criminal law was civil law. That is, there were no public prosecutors, no police, nothing analogous to that. And in 18th century England, which I've also written about, you had public courts, but criminal law was privately prosecuted and there were essentially nothing we would recognize as policemen. So that's one question is why I have the two. A second question is why allocate offenses the way we do? Why are certain things torched and certain things crimes? And a third question, which in a way undercuts my simple division between civil and criminal law is, why is civil law one bundle of legal rules and criminal law a different bundle? That is, if you start asking people, what are the defining characteristics of civil law? Or if criminal law, you get different answers from different people. And that's because each of them is really a bundle of rules which we're used to seeing together, but which can in fact be separated. So that again, if you think about 18th century English law, you have offenses which they call criminal. They are criminal in as much as the punishment is controlled by the state. But the enforcement, the prosecution in that system was private. So that's splitting out and giving you something that looks a little like our civil and a little bit like our criminal law. The part of the question is, why does each group of rules have the bundles it does, and perhaps are there other ways that you might be able to re-bundle and get a better legal system? What are the characteristics we typically associate with the criminal versus civil law? One of them is the question of who owns the claim against the offender. That in a civil system, the victim owns the claim. The victim can settle out of court if he likes for that claim. The victim can choose not to prosecute. In a criminal system, all of that applies to the state. To take sort of one neat example of the distinction, Blackstone writing in the 18th century describes a legal action called an appeal of felony, which he said is still on the books but nobody uses anymore, which was a private suit for a felony. And if you won, the guy who you sued got executed. And he says, if someone is convicted of an appeal of felony, the crown cannot pardon him, for the king is not a party to the suit any more than the king could dismiss the judgment in a civil case. So that sort of makes the distinction even in that extreme case. So one difference is who owns the claim. One difference is who prosecutes the claim, usually the same party but not necessarily. One difference is that we're used to associating criminal law with punishment and civil law with fines, that is damage payments, although you certainly get fines in the criminal law and you get injunctions, which are a little bit like punishments in the civil law. Another difference is the standard of proof is substantially higher in the criminal law than in the civil law. And another difference is that people think of criminal conviction as carrying a stigma that civil conviction does not. So that if you are annoyed at President Clinton, you might describe him as a criminal, but it probably would not occur to you to describe him as a tort-feasor. So in any case, I would say these are not, it isn't a natural distinction in the sense there are other ways of bundling it, but it's worth looking at the distinction as it exists in our law. And let me start out with what I regard as some wrong answers to the question of why we have civil law. One of them which is sort of embodied in, I mean why we have criminal law, civil is easy, why we have criminal law. And one wrong answer which is embodied in the way our legal rules are set up is that somehow the victim isn't really the victim. So that if somebody were to break into my house and steal my property, the resulting legal action would not be Friedman versus Smith, it would be the state of California versus Smith. And the idea, when people want to justify this as they say, well, robbery really scares everybody so it's a diffuse harm. And I think that's very implausible, at least as an explanation of the divisions we actually make because there are lots of torts, for example, people who run into my car. That's a much more diffused injury in the sense that it's much harder to connect the two, at least as diffuse. I'm just as scared, back I'm scared of driving down the highway than I am sitting in my house of being robbed. And that if you look at what things are torts and what things are crimes, you can't explain it by saying the crimes are the ones where there's a very spread out injury and the torts are ones where it's focused. And in fact we have rules and tort law designed to deal with spread out injuries. The standard one is the class action. There is of course a much better rule that we don't use and that is that tort claims could be transferable. This is one of the multitude of respects in which we have a less advanced legal system than the Icelanders did in the 10th century. That if you had a system in which tort claims were transferable, then if you're worried about lots of small torts, lots of torts that injure everybody by ten cents, I could sell to a firm that specialized in such things, all claims I had for minor torts that occurred this year. The firms could then assemble such bundles from millions of people and then when a particular case came up they could sort out all of the claims for a particular offense, sell them to an attorney who would litigate the case. So in fact there are perfectly good mechanisms under tort law for handling diffuse cases and sometimes we use some not so good mechanisms for the same purpose. Another argument is that the purpose of criminal law is to express public condemnation. But it's hard to see how that can be the reason because after all you could easily enough have the public court say for certain torts and by the way you're also a bad person and for other torts not say that. There's no particular reason why that requires the public-private enforcement distinction. There is another very interesting one which does give you a reason why some things should be torts and some things should be crimes but unfortunately the wrong things and that has to do with the incentives of the victim. That one of the many differences between the tort and a crime is that when I am the victim of a tort I acquire a valuable property right, namely the right to collect damages. The fact that when I am the victim of a tort I acquire a valuable property right reduces my incentive to avoid being a victim of a tort. And if you work through the analysis and it would take longer than I have to do you can divide offenses into two cases. One of them is something like a traffic accident where my taking precautions doesn't impose any cost on you reduces the chance of either as being in the accident and the optimal rule is that each person should act as if he bore 100% of the cost because that gives you the right marginal incentive if when I run into you I have to fix my car and pay for fixing your car then I will take any precaution which is cost effective in terms of reducing the probability of the accident, that is anyone with a reduction of probability times total loss is greater than my precaution cost. And if you have to pay my damages and fix your car you have the right incentives. So we get both of us the right incentives by saying that when we have an accident I pay a fine equal to your damages you pay a fine equal to my damages no one collects anything it all goes to the state and thus each of us is bearing 100% of the costs and thus each of us has the right marginal incentive to take precautions. On the other hand consider the case of burglary when I put a lock on my door I benefit myself of course but I do it at the expense of the burglar and the easiest way of thinking through the logic of this somewhat odd case is to imagine counter to reality a world where the cost of catching and punishing burglars was zero a world where you could always catch them and they could always pay a fine equal to damage done and in that world the only reason somebody would steal something was that it was more convenient than buying it or that you know I wasn't there to sell it you know Posner's locked cabin in the woods where a lost hunter breaks in because he's starving that sort of thing and in that world locks have no social benefit there's no reason to put a lock on my door because I'm going to get paid for the thing as soon as the police catch up with him if burglary is a tort I have no incentive to put a lock on my door in that world because I'm going to get as a damage payment the value of what he took but if burglary is a crime then the damage payment goes to the state so I have an incentive to engage in inefficient defensive precautions now generalizing this to the case the real world case where it's costly to catch and punish people makes the analysis a good deal more complicated you will find it in my Journal of Legal Studies article on efficient institutions for the private enforcement of law but it turns out that you can generalize it and that the basic rule is right that you want to use just from the victim standpoint you want to use criminal law to cover auto accidents and tort law to cover burglary and this does not do a good job of explaining the way our actual legal system sorts things into those two categories since we have it exactly backwards from that standpoint now of course it may occur to you that if we use the system I've just recommended for auto accidents then the first thing that happens when my car runs into your cars we both look around to make sure there are no cops and then drive off again because we have no incentive to report it and that of course raises another explanation which is that we use civil law when the easiest enforcement mechanism is to have the victim do the enforcing and that is certainly an argument for civil law that applies of course to burglaries as well that I'm the person who knows about it when I'm burgled and absent insurance there's very little incentive to call the cops and ask my time talking to them and I know they won't get the stuff back and if they arrest the person it doesn't mean oh good if they arrest them I spend more time in court also waste it so from an incentive standpoint there's much to be said for the civil rule of having it the victim who owns the claim but that doesn't explain why I have criminal law so it again doesn't sort the two out let me now go to what seems to be a couple of more interesting answers and answers that are at least in part true but one of the answers that Posner offers in his work is that you have criminal law because of the problem of the judgment proof defendant that is that under civil law my incentive to sue the person who is imposed a cost on me is to collect damages and if I'm sure he can't pay damages then I have no incentive to sue him so therefore judgment proof people can get away with offenses under a pure civil system because nobody will sue them and you then go on to say consider the case of offenses where it's hard to catch the person hard to detect who did it like burglary for example in those cases if you want the expected punishment to be equal to the damage done which is the sort of standard this would be the efficient rule in a world of costless enforcement you have to scale up the punishment to allow for the probability to be imposed so if you if the guy steals a thousand dollars and you only have one chance in ten of catching him you want to have a ten thousand dollar fine so if you scale up the punishment the less likely it is that he'll be able to pay it so that again suggests that the judgment proof defendant is at the root of the distinction and in particular in the case where it's hard to find him and catch him then the fact he's judgment proof is a serious problem because then it costs you money to catch him and you don't get the money paid back if it was very cheap to prosecute people you might prosecute them even on a small chance of being able to collect the damages and if you think about it this way of thinking about the problem does explain why several different features of tort law are combined I mean of criminal law are combined one of them is that I've just told you that we're considering cases where we have to engage in inefficient punishments the efficient punishment is the damage payment what you pay I get locking you up what you lose nobody gets for the prison that's a very inefficient punishment execution is a slightly less inefficient punishment what you lose nobody gets but at least we don't have to pay much to do it outside of Florida and that would be the subject of why not hang them all by the way why don't we shift to the more efficient hanging instead of the less efficient imprisonment and then things that develop out of that puzzle hanging with lower probability of course 1,000 of being hanged but that would be a digression in any case given that we're using inefficient punishments for criminal law that has several effects one which I won't go into is it changes the calculation of the optimal punishment because you have to ask not only do I want to deter this offense but how much is it worth spending to deter this offense because there's a net cost to doing it but also the fact that we're using inefficient punishments explains why we have a higher standard of proof because a high standard of proof reduces the chance of punishing innocent people if the punishment is a costless punishment, if punishing you means you pay me a thousand dollars in damage that you really didn't owe me well so you occasionally punish the innocent person, occasionally I pay some money I didn't know, occasionally someone pays me some money he didn't owe me all comes out in the wash but if punishing means executing somebody who didn't commit a crime occasionally I get executed for a crime I didn't do, occasionally you get executed for a crime you didn't do that doesn't balance out alright it's an expensive punishment so therefore where you're using expensive punishments there's clearly a good efficiency argument for being more careful to make sure you're imposing them on the right people and that is one of the respects in which our law does get it right furthermore I would claim that the fact we're using efficient punishments also explains the moral sanctions that if you think very very approximately of tort law as a pricing system where I impose some risks on you and I pay for them in a damage payment then we don't want me to have a higher cost than that we don't want to have me to feel bad about it as well right it's an efficient action but in the case of murder it isn't that we upset the punishment so that any murder that isn't deterred is a good thing alright this is if you do too much reading in law and economics without thinking while you're doing it you say well look once we've got an optimal system the punishment for murder is equal to the damage done so the only murderers we don't deter are efficient murders and the reason we don't raise the punishment is we're afraid we wouldn't have enough murders that's wrong that would be right in a world where it costs nothing to catch and punish people because in that world you really would raise the punishment that high but in the real world where catching and punishing are costly you raise your expected punishment to the point where deterring one more murder isn't worth the extra police enforcement it would take to do it at which point the murder is still doing a large net damage and therefore you would like to call in moral sanctions as well as an additional dimension of punishment as it were to try to prevent things that are in fact on net costly so all of that gets explained let me give you however another argument and this is an ingenious argument due to Landis and Posner who unfortunately have not correctly followed it through or entirely correctly understood it in the original article and this is the following problem suppose we have a some system for private enforcement of criminal law and we've got an industry of private enforcers and the private enforcers somehow get a claim against a criminal maybe they buy it from the victim and then when they can bring him in and prove he's guilty they get to collect the fine that he pays and this is a system that is sketched out by Becker and Stigler in a famous article and Landis and Posner wrote a response to that and Landis and Posner made the following very ingenious point that the optimal system for enforcing criminal law we have two variables we want to optimize the probability of catching an offender and the punishment if we do catch it and if you think about it as an economist you could say we can work out a total cost function for producing deterrence starting with a production function our two inputs are effort catching criminals which gives us the probability and costs born punishing them which gives them as a punishment consider all of the probability punishment payers which have the same effect which are equivalent from the standpoint of the offender high punishment low probability, low probability high punishment pick the probability punishment payer that minimizes the sum of enforcement and punishment cost where punishment cost includes the cost to the person being punished as well as to the rest of us that is the cheapest way of producing that level of deterrence you do it for every level of deterrence you have a total cost function so that tells us that an optimal system wants to set the right level of probability and the right level of punishment aha Landis and Posner say but under your private enforcement system the fine paid by the criminal is simultaneously the punishment to him and the reward to his enforcer the reward to his enforcer will determine how hard it's worth firms working to try to catch these people so when you have sent what the punishment is for an offence you've simultaneously set both probability and punishment and with a single control variable you can't separately optimize two different things alright people follow it it's a clever argument it just demolishes tort law as well as criminal laws it turns out but it's a clever argument and so he says at least an ideal governmental system could separately set the probability how many police they have and the punishment what they do to people and could thus get the double optimization problem and that's a need argument the reason that it turns out though they didn't realize it to be an argument against tort law as well is that in tort law the damage punishment the damage payment paid by the tort is simultaneously the punishment to him for sending it to the victim to litigate well you can spend more or less money on litigation according to how important it is to win there is if one only new enough an optimal expenditure on litigation designed to reduce errors and an optimum punishment and there is no reason to think that by setting one number namely the damage payment you can generate you can simultaneously optimize both numbers so it turns out to be a problem both ways it is however a problem that I have solved because I wrote an article in response to Landis and Posner and it turns out that there is a rather neat solution it doesn't apply to everything but it applies to a lot and the solution works as follows let the court set not the punishment but the expected punishment the average punishment and suppose they say that for a particular class of offense the expected punishment should be a thousand dollars I have a private enforcement firm has bought a thousand claims the claims from the victims of a thousand robberies I am allowed to impose an average punishment of a thousand dollars on those thousand offenders well I can do it by catching all a thousand and finding them a thousand dollars I can do it by catching one and hanging him assuming that's the equivalent of a million dollar punishment or anything in between think of any punishment as consisting of a punishment imposed and a punishment collected in the case of a fine you pay a thousand dollars I collect a thousand dollars in the case of execution you pay a life I collect zero in the case of imprisonment you pay your liberty I collect minus the cost of running a jail system punishment cost is just the difference between fine paid and fine collected in my system the private enforcement agencies are allowed to impose a given value of expected fine paid effective punishment paid the punishment from the criminal standpoint they then are collecting the punishment collected so what does that mean it means that if they catch all thousand offenders and find them a thousand dollars and they can pay a thousand dollars then the enforcement agency collects a million dollars if they catch one guy and hang him they can collect nothing that means that punishment cost is a subtraction from their revenue that is the total punishment imposed has been fixed by the court because the court fixed the expected value of punishment imposed so I know that having bought a thousand offenses I can impose on the criminals a total cost of a million dollars how much of that I can collect depends on whether I have to use inefficient or efficient punishments so I can opt to spend the money for all thousand of them and then I can actually collect my thousand dollar fines I'm paying a high enforcement cost and a low punishment cost or I can opt to just find the one guy who was stupid enough to do his burglary while a photographer was passing by and I hang him but I don't get any money work through the logic of that problem you can find the article on my web page and it turns out that the sum of enforcement cost plus punishment cost is a subtraction from the firm's revenue and that the firm therefore in order to maximize its profit picks the probability punishment pair which minimizes that sum which means it's just done one of the maximizations automatically right the way we pick the optimal punishment probability pair is to minimize the sum of punishment cost plus enforcement cost those are both private costs for the enforcement firm so it is in the private interest of the enforcement firm to you pick the optimal probability punishment pair and we've just made half of our maximization endogenous to the market and we are now left with the other half which is the court has to set the right expected punishment and we could discuss in another talk how we do away with that half but for the moment I'm assuming that the government is still running the courts now there is however a problem with my solution and the problem really gets us back to the judgment proof defendant problem the problem with my solution is what happens if fine collected is less than enforcement cost what happens in other words if the net cost to the private enforcer of catching somebody is greater than the amount of money they get out of it which is not unlikely for the standard criminal case you may not be able to pay a fine and the answer is if you just sort of run through the equations as a good economist well that's just fine that just means the price at which they buy offenses is negative alright but you say wait a minute why would I as the victim pay the enforcement agency to take my case why don't I just ignore the whole issue and say well I got robbed of luck the answer to that of course is that the reason I will pay the enforcement agency to take the case is that I pay them before the robbery happens not after that what I am doing is signing a contract with the enforcement agency by which they agree that if anybody robs me they will impose the court mandated expected punish they will buy the claim and on their collection on their portfolio of claims they will impose the court mandated expected punishment and of course I then put a note on my door which is a copy of my contract with them for any would be burglars to read or in other words what I have just done is to convert deterrence into a private good now for those of you who think this is a wild eyed idea that only a crazy 20th century anarcho-capitalist would think of the English were doing this routinely in the 18th century the equivalent of this under 18th century English law was called an association for the prosecution of felons this is a system where there was no private incentive to prosecute felons in terms of collecting damage payments because felony was a crime so when somebody robbed you and you spent time and effort getting him convicted and someone hanged him or possibly transported him nonetheless there were private prosecutions and one of the reasons is that a bunch of people in one town who think they are likely victims of crime form an association for the prosecution of felons each of them chips in some money to a pooled fund which will therefore be spent prosecuting felons felonies committed against any of the members and final crucial step the list of members of the association is published in the newspaper for the felons to read that was in fact a real world system and there were thousands of these associations in England for making deterrence of private good so it looks as though I've now solved quite a lot of both haves of the problem that is both the double maximization problem and the judgment-proof defendant problem if we have private enforcers it's a tort system in that the claim belongs to the victim it's a tort system in that he can transfer it of course to anybody else as any sensible tort system would let him do unlike ours and the reason why he transfers it at a negative price in the case of crimes where the expected return from catching the criminal is negative we are still left however with a hole in this system and the hole in this system is the problem of anonymous offenses by anonymous offenses I don't mean offenses where the offender is anonymous that's an equal problem for all systems you have to find him, I mean ones where the victim is anonymous that in order for deterrence as a private good to work the criminal has to know who is committing the offense against him so that you need some equivalent of the stamp on the door and he knows whether or not you are protected whether or not you have contracted to have somebody catch criminals who commit offenses against you and therefore is deterred and in the English case in the early 18th century they were concerned about inadequate incentives to prosecute and they were particularly concerned with that with regard to certain offenses for example highway robbery because highway if somebody breaks into my house and I am a well off English middle class person who is a member of a prosecution association the person who breaks into my house lives in the same town I do he has prudently checked the newspaper to see if I am a member of prosecution association and will be suitably deterred but when a highwayman stops me in the open road he doesn't know enough about me to know whether or not I have arranged to prosecute him hence I have too low an incentive to do it hence the mechanism doesn't work very well for that there were problems with the reward system which again would get me into the other talk which you can bring me to give some other time but nonetheless that's the hole in the system the hole in the system is that it doesn't cover cases where both the market price of the offense is negative that is to say where you can't collect as much as it costs to collect it and where the victim is anonymous meaning that you can't deter the criminal by in some way making it possible for him to know whether or not the victim is protected and there are going to be some such offenses so I don't have a my answer to should we abolish the criminal law isn't quite an unambiguous yes because there is an area where a sufficiently well run criminal system would improve on what the private system could do but we don't of course have the option of a sufficiently well run criminal system so it's now unclear whether in fact we would be better off off abolishing it but notice that this also brings us back to the distinction I made in a comment earlier between a locally and a globally efficient system and back to the sort of long discussion that Posner set off 27 years ago of whether or not the common law is efficient because what I am suggesting is that given the general way our system is set up it indeed does make some sense to have robbery and similar offenses be crimes there are some arguments against it but there is a real argument for it due to the judgment proof defendant problem that in order to get around that you have to substantially re-engineer the system you have to create a system where the first you have to put in probability multipliers for tort law which we don't normally do with some rare exceptions you have to put in a system where the court sets the expected punishment rather than the actual punishment and all of the rest of this so I am suggesting that one can in fact get an improved system but that the detailed features of our system things like higher standard of proof for criminal and moral sanction for criminal do indeed make a certain amount of local sense from an efficiency standpoint how much time do I have left oh wonderful I didn't realize I was talking that fast I hope that they are taping this a little bit faster than the usual speed so that they can play it back at a delayed speed given that let me say let me give you a very brief version of the other talk that I'm not giving and that starts out with a very simple puzzle and the puzzle comes out of my probability punishment point before the idea that you can think of a production function for deterrence suppose we have an offense which is currently punished by catching ten percent of the but catching sixty percent of the offenders and putting them in jail for ten years and suppose that after a little market research we discover that from the standpoint of an offender a sixty percent chance of being put in jail for ten years is exactly equivalent to a ten percent chance of being executed well we can clearly improve this system we continue to catch sixty percent of the offenders and when we catch one of them we roll a die one through five we turn them loose six we hang them this system is identical ex-ante from the standpoint of the offenders they are facing the same expected cost through punishment it therefore provides the same deterrence so it's identical from the standpoint of the victims and it saves us the cost of jails and obviously we can make a further improvement in this system there's no point in catching people and turning them loose so we just fire police until we're down to only catching ten percent of the offenders we hang them all those of you who know something about 18th century English law will understand why I first got intrigued by that system because that was a system where practically all major offenses were capital although not everybody was hanged but they were all capital offenses on paper in any case that's why why not hang them all and you can then the argument as I've given it is right it's right but it shows something I'm going to argue wrong with the structure of current law and economics the next step of course is to say we can make the system even more efficient than that for one thing there's no point in throwing the bodies away when we hang people because the ideal situation for transplanting organs is when the donor is still alive so therefore the capital offenses you will all forfeit your organs to the state for transplantation and there are a variety of other ways one could improve the efficiency of the punishment system for example penal slavery systems in which people are told you will have a probability of being executed unless you pay a fine of course and you can reduce the probability of being executed by paying a fine so that you push people into making sure they're not judgment proof you'll notice that whether people are judgment proof itself is not an exogenous variable the same offender who is judgment proof in our system against a fine for robbery I will predict as not judgment proof when somebody comes to collect the assembling debts from him because they have more effective enforcement mechanisms than we do so you could in fact if you start thinking about it you could redesign our punishment system in a form that would be much more efficient that in terms of our standard analysis is provably superior to what we now have and that we would all think was terrible and the question is is there something wrong with doing that and the answer is yes what's wrong with this system without the whole system done we have a nice efficient system for punishment which means that when I convict Mario of an offense he loses an eye game and that gives me an incentive to convict him of an offense whether or not he's done it I started thinking about this problem after an argument on the internet many years ago when I was proposing a system of a private enforcement system where people worked off their debts as penal slavery essentially and somebody said yes the colonial powers used that system in Africa when they wanted a road built they found some excuse to arrest a bunch of people and set them to building roads so that in fact the problem with an efficient system is the incentives for the enforcers that a system in which what the criminal loses someone else gains is a system in which if I can get the court system on my side I can't appropriate your economic, physical and biological capital alright human, physical and biological capital that doesn't sound like a very good very good system or at least it sounds like a potentially very expensive rent seeking struggle now this is not a purely imaginary problem either very few of my problems are purely imaginary if you're willing to look at enough historical examples and the historical example here is very easy to find because it's all around us the two examples of this problem that we are all familiar with are civil forfeiture and punitive damages that after all civil forfeiture is this ingenious system by which when property is used in a criminal fashion whether or not the owner was responsible the property forfeits to the police to the state but in fact to the police department the way it's usually done and it's obvious if you just think about it that's an improvement because that's a more efficient punishment to pull up but then you think a little longer and you say well in that case if I'm a rational police officer what do I try to maximize well I try to make sure that when my paid informer does a drug deal he does it on the property of a nice expensive house alright and there is one notorious case in Monterey where a group of law enforcement agencies came on to the property of an eccentric wealthy man who owned in a state worth some millions of dollars looking for marijuana they didn't find any marijuana they came into his house he came down the stairs and they shot him dead and I gather from the report by one of the local prosecutors of course the people who did the killing never got prosecuted but the case did get discussed that the evidence was that before coming on to his property the research they did was not on whether he was growing marijuana but on the market value of his estate somebody who they reasonably thought might be growing marijuana and if he was the estate forfeited to them meaning to the organization they were a part of so that's a case of precisely the rent seeking problem that because it is profitable to get property to forfeit you have an incentive to try to get it to forfeit rather than to try to enforce the particular laws punitive damages the so-called liability crisis and so forth I don't know enough to be sure if it's really as bad as some people say that you can see the same argument that tort damages are a very efficient punishment and therefore you have to worry about the incentives and that gives me to one final point which is again a case where the system may be locally efficient though perhaps not globally and that's the question of why we don't have probability multipliers in tort law punitive damages may be a special case but punitive damages didn't matter until about 40 years ago in the 19th century it was the opinion of the Supreme Court of New Hampshire that there was no such thing as punitive damages in the Anglo-American legal system and that was an influential minority opinion at the time they weren't an important issue why is the standard tort rule that you get enough to make the victim whole rather than enough to actuarially make them whole if you only have a 50% chance of catching the tort freezer why don't you get twice the damages done when you catch them and my answer is that doing that would be an invitation to fraudulent tort that if when you run into my car and I catch you, I'm a great deal richer than if you hadn't run it into it in the first place which it would be if hit and run is say multiply by 10 injury if we only catch a tenth of the hit and run people then I have a strong incentive as both the victim and the prosecutor in a civil system to arrange to shove my car into your way when four of my friends who will lie for me are watching think about it in terms of insurance we all study moral hazard and it occurs to some of us that moral hazard has an odd discontinuity if I insure my factory for 80% of its value the chance it burns down is a little bit higher than if it wasn't insured because I don't take as many precautions if I insure it for 90% of its value the chance it will burn down is just a little bit higher if I insure it for 110% of its value suddenly the chance it will burn down gets very very very high there's a discontinuity in that system well I suggest that one reason why traditional tort law has been reluctant to include multipliers and why maybe under the law is it now punitive damages might be a mistake is that you've got that same discontinuity that once having the accident occurs actually makes you better off if you've set up the situation there is then a serious risk of fraudulent for fraudulent accidents and that's a problem which I ought to incorporate in my broader subject but since you've just gotten two talks for the price of one I think I'll stop