Previous Chapter

Chapter 20

The Economics of Law and Law Breaking


This chapter consists of five parts. The first two take the existing legal structure as given, using economics first to understand criminal activity and suggest ways of defending against it and then to analyze the net costs of crime and in the process to suggest why certain things should be illegal.

The remaining parts of the chapter take the existing system of laws and law enforcement as an object of study rather than a framework within which to study crime. The third part sketches the analysis of what the punishment for crimes ought to be and how hard we should try to apprehend criminals: how much we should spend and what fraction of the criminals we should catch. The fourth part discusses the advantages and disadvantages of two alternative systems for catching and convicting offenders--public, as in our present criminal law, and private, as in our present civil law. The final part of the chapter works through, in some detail, the economic analysis of the civil law of accidents.



By the economics of crime, I do not mean the effect of crime on the GNP or why poverty causes crime. Economics means the same thing here that it meant in Chapter 1 and has meant throughout the text--a particular way of understanding human behavior. The economic approach to analyzing crime starts from the assumption that a burglar burgles for the same reason I teach economics--because he finds it a more attractive profession than any other. It draws the obvious conclusion that if one wishes to reduce burglary--whether one is a legislator or a homeowner--one does so by raising the costs of the burglar's profession or reducing its benefits.

Many years ago, I had a disagreement with a friend concerning a practical element of life in New York City; our argument illustrates in a small way the difference between the economic and the noneconomic approach to crime and criminals. I was at the time living in a somewhat hazardous part of Manhattan. When I found it necessary to go out at night, I was in the habit of carrying with me a four-foot walking stick. My friend argued that I was making a dangerous mistake; potential muggers would feel challenged and swarm all over me. I felt, on the contrary, that muggers, like other profit-maximizing businessmen, would prefer to obtain their income at the lowest possible cost. By carrying a stick, I was not only raising the cost I could inflict on them if I chose to resist, I was also announcing my intention of resisting; they would rationally choose easier prey.

I never got mugged, which is some evidence that my view of the matter was correct. More evidence comes from observing who the people are who do get mugged. If muggers are out to prove their machismo, they ought to pick on football players; there is not much glory in mugging little old ladies. If muggers are rational businessmen seeking to obtain revenue at the lowest possible cost, on the other hand, mugging little old ladies makes a lot of sense. Little old ladies--and other relatively defenseless people--get mugged. Football players do not. It is said that someone once asked Willie Sutton why he robbed banks. "That's where the money is" was his (rational) reply.

The same analysis that helped me decide what to take with me on my evening strolls around Manhattan's Upper West Side can also be used in analyzing a question that often comes up in discussions of gun control. Opponents argue that gun control, by disarming potential victims, makes it more difficult for them to protect themselves from criminals. Supporters reply that since criminals are more likely to know how to use guns than victims, the odds in any armed confrontation are with the criminal. This is probably true, but it is almost entirely irrelevant to the argument.

Suppose one little old lady in ten carries a gun. Suppose that one in ten of those, if attacked by a mugger, will succeed in killing the mugger instead of being killed by him--or shooting herself in the foot. On average, the mugger is much more likely to win the encounter than the little old lady. But--also on average--every hundred muggings produce one dead mugger. At those odds, mugging is a very unattractive profession--not many little old ladies carry enough money in their purses to justify one chance in a hundred of being killed getting it. The number of muggers--and muggings--declines drastically, not because all of the muggers have been killed but because they have, rationally, sought safer professions.

When, as children, we first learn about the different sorts of animals, we are likely to imagine them arranged in a strict hierarchy, with the stronger and more ferocious ones preying on everything below them. That is not how it works. A lion could, no doubt, be fairly confident of defeating a leopard, or a wolf of killing a fox. But a lion that made a habit of preying on leopards would not survive very long. While any particular leopard would probably lose the fight, there would be some small chance of the lion getting killed in the process and a larger chance of his getting injured. That is too high a price for one dinner. That is why lions prey on zebras instead. In just the same way, a potential victim does not have to be more deadly than the criminal, just deadly enough so that the cost to the criminal is a little greater than the benefit.

This does not, of course, prove that gun control is a bad thing; while I have rebutted one argument, there are many others, both pro and con. It does illustrate an important general principle. In analyzing conflict, whether between two animals, criminal and victim, competing firms, or warring nations, our natural tendency is to imagine an all-out battle in which all that matters is victory or defeat. That is rarely if ever the case. In the conflict between the mugger and the little old lady, the mugger, on average, wins. But the cost of the conflict--one chance in a hundred of being killed--is high enough so that the mugger prefers to avoid it. In this case, as in many others, the problem faced by the potential victim is not how to defeat the aggressor but only how to make aggression unprofitable.


Economics Joke #3: An economist and a businessman were walking in the woods when they encountered a hungry bear. The economist turned to run. "That just goes to show how ridiculous you economists are with your assumptions," said the businessman. "You're assuming you can outrun the bear." "Wrong!" replied the economist. "I'm only assuming that I can outrun you."


Economics of the Spaceways

There is a nice fictional illustration of this point in a science fiction story by Poul Anderson. The setting is a far future in which interstellar travel and trade are common. There is a potentially profitable trade route connecting two groups of stars. Unfortunately the route runs through the territory of a nasty little interstellar empire. The nasty little empire ("Borthu") has the unpleasant habit of seizing passing starships, confiscating their cargo, and brainwashing their crews; the crew is then added to Borthu's fleet, which is critically short of trained manpower.

Borthu is a nasty little empire; the trading corporations could, if they chose, get together, build warships, and defeat it. But doing so would cost more than all of the profits to be made on the trade route. They could also arm their trading ships--but the cost of building and manning an armed ship would more than wipe out the profit the ship would generate. They can win--but, being rational profit maximizers, they won't.

The problem is solved by Nicholas Van Rijn, the head of one of the trading corporations--after he has first persuaded his competitors to agree to pay a fraction of their profits on the route to whoever solves the problem. The solution is to arm one ship in four, reducing the profit but not eliminating it. Warships carry larger crews than merchant ships. Three times out of four, the empire attacks a trading ship, capturing it and its four-man crew. One time out of four, the trading ship is armed; the empire loses a warship and its twenty-man crew. Every four attacks cost the empire, on net, eight crewmen. Piracy is no longer profitable, so it stops.

The logic of the problem, and the solution, is nicely summed up in Van Rijn's reply to one of his colleagues, who suggests that they should fight even if it costs more than the trade is worth to them.

"Revenge and destruction are un-Christian thoughts. Also, they will not pay very well, since it is hard to sell anything to a corpse. The problem is to find some means within our resources to make it unprofitable for Borthu to raid us. Not being stupid heads, they will then stop raiding and we can maybe later do business."

--"Margin of Profit," in Un-man and Other Novellas by Poul Anderson



I recently came across another discussion of the economics of crime and crime prevention in a book on the subject written from the inside--in several senses. The title was Secrets of a Superthief (by Jack Maclean). The author, according to both his own account and the journalist who wrote the introduction, was a spectacularly successful burglar, specializing in high-income neighborhoods. As he tells it, he ran a class act--when a house contained nothing he thought worth stealing, he would pile up the rejected booty on the kitchen table and steal the control panel from the burglar alarm. His general policy was to reset burglar alarms on his way out, to make sure no less discriminating thief broke in and messed up the house.

Eventually Superthief made a small professional error and found himself taking an unplanned vacation, courtesy of the State of Florida. Being an energetic fellow, he spent his time behind bars polling fellow inmates on their techniques and opinions and writing a book on how not to get burgled. One of Superthief's principal insights is the same as Van Rijn's--the essential objective in any conflict is neither to defeat your enemy nor to make it impossible for him to defeat you but merely to make it no longer in his interest to do whatever it is that you object to.

In giving advice to potential victims, Superthief argues that making it impossible for a burglar to get into your house is usually not an option; few doors will stand up to a determined burglar properly equipped.The function of strong doors and locks is not to make burglary impossible but to make it more expensive, by increasing the skill and equipment needed by the burglar as well as the chance that he will be detected before finishing the job.

A less expensive approach is to use what Superthief calls "mind games." Figure 20-1a shows my version of one of his suggested tricks--in the form of a note taped to my back door. Both Mrs. Jones and Rommel are wholly imaginary. A potential burglar may suspect that, but he has no way of being sure. Exterminators are common enough in this part of the country, the reference to the back rooms is vague enough to make it uncertain just where he can go without breathing insecticide, and Rommel, presumably a German shepherd or Doberman pinscher (can you imagine a poodle named Rommel?), is in the room that, according to Superthief, burglars consider most worth robbing. Superthief's version of the note referred to pet rattlesnakes loose in the house--a better story than mine but less likely to be believed. Superthief gives many other examples of simple and inexpensive mind games--such as leaving a large dog-feeding dish or a jumbo-sized rubber bone lying around the backyard.

Figure 20-1

Low-cost burglar repellents. Fictitious notes to a fictitious cleaning lady and a real son.

Figure 20-1b shows another of my precautions--a solution to a problem that Superthief does not discuss. One of the rooms in the back part of my house contains some things that a thief might well find worth stealing; for that reason, I have equipped it with its own deadbolt lock. This raises a problem. A rational thief will assume I am a rational victim and deduce that if I have a lock on that door, it is probably because I have things worth stealing behind it. My solution is the sign shown in Figure 20-1b. It is intended to suggest an alternative explanation for the lock--dangerous chemicals in the room and a curious child in the house. The solution is original with me, but I believe Superthief would approve.


Illegal Markets

"(On earth they) even have laws for private matters such as contracts. Really. If a man's word isn't any good, who would contract with him? Doesn't he have reputation?"

--Manny in The Moon is a Harsh Mistress by Robert Heinlein


So far, we have been using the economic analysis of crime to figure out how, on an individual level, to deal with it; the discussion is in that sense practical as well as theoretical. Before ending the section and going on to analyze questions of law and law enforcement, we will first use economic analysis to explore an equally interesting but less immediately useful question--the structure of illegal markets.

We are used to thinking of markets as public, socially accepted institutions such as the stock market, the wheat market, or a supermarket; one important feature of such markets is that the agreements to buy and sell made in them can usually be enforced, if necessary, in the courts. But the concept of a market is much broader than that. There are markets for political influence in the Soviet Union--and in Washington. There are markets for illegal drugs and stolen goods. There are markets for sex, both legal (see Chapter 21) and illegal. The enforceability of contracts by public courts may be useful for the working of markets, but it is certainly not essential.

Economics applies to illegal markets as well as to legal ones. When one input to production is eliminated, substitutes become more valuable; if contracts cannot be enforced in the courts, alternative ways of getting people to abide by their contracts become more important. We would expect substitutes for the service provided by the court system to be important elements of illegal markets. One substitute is reputation. The traditional definition of an honest politician is one who stays bought. Another and more violent substitute for the courts will be discussed a little later.

Another characteristic of illegal markets is that the handling of information is more costly than in legal markets; the same information about your employees that you want in order to decide on your future dealings with them is also useful to a prosecutor deciding on his future dealings with you. This is one of the reasons that I (and others) regard "organized crime" or the "Mafia" as largely mythical, at least as they are commonly portrayed. A "General Motors of Crime" makes little sense. Someone at the top of such a firm has to know what people at the bottom are doing, in order to (among other things) decide whether they are earning their pay. Passing such information up many levels of hierarchy would be hazardous in the extreme. It seems more likely that most crimes are committed by individuals or small firms and that organized crime is analogous, not to a giant corporation, but to something more like a chamber of commerce or better business bureau for the criminal market.

Such an interpretation flies in the face of what we are usually told, in newspapers and congressional hearings alike. Before you reject it on that basis, it is worth considering how credible the sources of information for the newspapers and the committees are and whether their incentives are geared to generating accurate research or exciting stories. Newspapers want to sell copies and politicians want to get reelected; announcing that organized crime is not a major threat seems a poor way of doing either. The sources of information are either law enforcement officials, who want to prove that they need more money and power to fight organized crime, or criminals testifying in exchange for immunity--with an obvious incentive to say whatever their captors wish to hear. It is interesting, in reading such accounts, to compare descriptions of the power and importance of the Mafia with descriptions of how the witnesses actually ran their criminal enterprises; the latter generally portray the witnesses as independent entrepreneurs, not employees of some criminal superfirm.

Academic studies of the criminal market involve certain difficulties not present in most other fields of research. Nonetheless, such studies have occasionally been done, and there is at least some scholarly evidence that seems to support my conclusions. For example, a study of illegal gambling in New York, based on records produced by police wiretapping, found that bookies were small independent operators and that numbers operators were somewhat larger but also competitive. Neither bookies nor numbers operators seemed to have much ability to use violence against their competitors; they even had difficulty enforcing profit-sharing agreements with the subcontractors who brought in their customers.

A more entertaining (but possibly less reliable) way of learning about organized crime is by reading what claim to be inside accounts. The Last Testament of Lucky Luciano contains a revealing incident. After a gangland war over who was to be Capo di Tutti Capi--boss of the Mafia--the winner called together gangland leaders from all over the country. He announced that:

everything would now be combined into a single organization under one rule--his. . . The key was discipline, Maranzano emphasized repeatedly, rigid discipline, with Maranzano himself the supreme arbiter of all disputes, as he would be supreme in everything. That discipline ... would be strictly enforced.

In less than five months he was dead.

My own conjecture is that what the Mafia really is, at least in part, is a substitute for the court system; its function is to legitimize the use of force. To see how that might work, imagine that you are engaged in some criminal enterprise and one of your associates pockets your share of the take. Your obvious response is to have him killed--murder is one of the products sold on the market you are operating in. The problem with that is that if people who work with you get killed and it becomes known that you are responsible, other participants in the illegal marketplace will become reluctant to do business with you.

The solution is to go to some organization with a reputation, within the criminal market, for fairness. You present the evidence of your partner's guilt, invite him to defend himself, and then ask the "court" to rule that he is the guilty party. If it does so--and he refuses to pay you appropriate damages--you hire someone to kill him; since everyone now knows that he was in the wrong, the only people afraid to do business with you will be those planning to swindle you.

That, I suspect, is one of the functions that the Mafia and similar organizations serve on the criminal market. This is a conjecture about organized crime, not something I can prove; but it is not, so far as I know, an implausible one.



So far, I have taken the legal structure as given and used economics to analyze the behavior of criminals. The next step is to use economics to analyze the cost imposed by crime. The objective in doing so is in part to show why, from the perspective of economics, certain things should be illegal, and in part to show how the analysis of the market for crime can be fitted into the general framework of economics. Here and in the remaining parts of this chapter, we will continue to assume that the participants in the criminal marketplace--criminals, victims, and law enforcers--are rational, correctly choosing the means to achieve their differing objectives.


What Is Wrong with Robbery Anyway?

We take it for granted that certain activities, such as robbery, theft, and murder, are bad things that ought to be prevented. From the standpoint of economic efficiency, it is not immediately obvious why. Theft appears to be merely a transfer; I lose $100 and the thief gains $100. From the standpoint of efficiency, that appears to be a wash--costs measured in dollars just balance benefits in dollars. If so, what is wrong with theft?

If that were all that happened, theft would indeed be neutral from the standpoint of efficiency, neither an improvement nor a worsening. It is not. Theft is not costless; the thief must spend money, time, and effort buying tools, casing the house, breaking in, and so forth. How much time and effort? To answer that question, we do not have to find any actual thieves and interrogate them. Economic theory tells us what the cost will be--at least for the marginal thief. In equilibrium, on the thieves' market just as on other competitive markets, marginal cost equals average cost equals price. The marginal thief has no gain to balance the cost he imposes on his victim. The analysis goes as follows.

Suppose that anyone who wished to become a thief could steal $100 at a net cost, including operating expenses, value of time, and risk of being caught, of only $50. Revenue is greater than cost, so economic profit is positive; firms enter the industry. If stealing pays better than alternative occupations, people will leave those occupations to become thieves.

As more people become thieves, the marginal return from theft falls. Many of the most valuable and easily stolen objects have already been stolen. Every diamond necklace has three jewel thieves pursuing it. A thief breaks into a house only to discover that Superthief has stolen all the more valuable jewelry--and reset the alarm. Just as in other industries, increased output drives down the return, although not for quite the same reason. The "price" that a thief gets for his work, the amount he can steal for each hour of his own time that he spends stealing, falls.

How far does it fall? As long as stealing yields a higher return than alternative occupations, people will leave those occupations to become thieves. Equilibrium is reached when, for the marginal thief, his new profession is only infinitesimally better than his old--and for the next person who considers becoming a thief and decides not to, it is infinitesimally worse. In equilibrium, the marginal thief is giving up a job that paid him, say, $6/hour in order to make, after subtracting expenses of his new profession such as lawyer's fees and occasional unpaid vacations, $6.01/hour.

So in equilibrium, theft is not a transfer at all. The marginal thief who steals $100 spends about $100 in time and money to do so. His costs and his return almost exactly cancel, leaving the cost to the victim as a net loss. The transaction is not a wash but a Marshall worsening of about $100.

So far, I have discussed only the marginal thief. What about the individual who is exceptionally talented at stealing or exceptionally bad at alternative professions, so that being a thief is more attractive to him, relative to other professions, than to the individual who just barely decided to become a thief? His situation is like that of the firm with a particularly good production function, discussed in Chapter 9; at a price at which marginal firms just cover their cost, the inframarginal firm, or thief, makes a profit. When he steals $100, he does so at a cost of only $50, leaving him $50 ahead. Since the victim ends up $100 behind, the result is still a Marshall worsening, although not by as much as in the case of the marginal thief.

So far, we have the following result. If all thieves are marginal thieves--if, in other words, there is not much variation among potential thieves in their comparative advantage for thievery--the net cost of theft, including costs and benefits to both thieves and victims, is about equal to the amount stolen. If thieves vary widely, the net cost is still positive, but less than the amount stolen.

There are two directions we can go from here in analyzing the cost of theft. One is to make the analysis more realistic by including some costs that so far have been omitted--the costs of defense against theft. These would include both private costs--locks, burglar alarms, security guards, and the like--and the public costs of police, courts, and prisons. While I have not actually done such calculations, my guess is that such costs are much larger than the net gains of theft to the inframarginal thieves, making the total cost of theft more, not less, than the total value of all goods stolen.

The other thing we can do at this point is to see how the analysis of theft fits into the general structure of economic theory. We will start by converting the verbal analysis of marginal and inframarginal thieves into something very similar to a conventional supply and demand diagram. We will go on to show that the undesirability of theft can be seen as merely a special case of something we already know--indeed of two different things we already know.

Figure 20-2 is a supply and average revenue diagram for theft. The horizontal axis shows hours of theft per year. The vertical axis shows dollars stolen per hour--the "wage" that thieves receive. The supply curve S, like any other supply curve, shows how much labor will be supplied at any wage--how the number of hours spent stealing depends on the number of dollars per hour that can be stolen. The average revenue curve AR shows how the revenue from an hour spent stealing varies with the amount of theft. As the number of thieves stealing increases, the return per hour falls, so AR slopes down, just like a demand curve.

By using the number of hours as my measure of quantity, I have implicitly assumed that the difference among thieves is in how much it costs them to spend an hour stealing, not in how much they can steal in an hour. Such differing costs reflect differences among the thieves in their taste for leisure, their earning opportunities in alternative employments, and their estimate of how likely they are to be caught. I could, if I preferred, take account of differing abilities to steal by defining my unit as some kind of standard hour--where an hour spent by an incompetent thief counts as half a standard hour, and an hour spent by Superthief counts as ten. Dropping this assumption would make the analysis a little more complicated without changing anything essential, which is why I am not doing it that way.

The market for theft. The shaded area L shows the net loss resulting from theft--the loss to the victims minus the producer surplus (colored) received by the thieves.

The curve labeled S is a supply curve, but AR is not a demand curve, although it serves much the same function in our analysis. If, in an hour, you can bake 20 cookies and sell them to me at $0.25/cookie, that tells us not only something about your opportunities but also something about my tastes. The fact that in 10 hours you can steal $50 from me tells us something about your opportunities to steal but does not imply that I like being stolen from, since the transaction is not a voluntary one. AR is not, like a real demand curve, equal to a marginal value curve. The area above S and below P is producer surplus, just as with an ordinary supply and demand diagram, but the area below AR and above P is not consumer surplus.

The total amount stolen per year is average revenue--the amount stolen per hour--times the number of hours of theft per year. On the figure, that is the shaded area plus the colored area. The thieves receive that as income, bear costs equal to the shaded area, and receive a producer surplus equal to the colored area. The victims lose the amount stolen and receive nothing. The net loss L, the shaded area under the supply curve, is equal to the loss to the victims minus the gain to the thieves. If S were nearly horizontal, corresponding to a highly elastic supply of thieves, net loss would be almost the entire amount stolen. That is the case described earlier, where all thieves are marginal thieves.

Supply and revenue curves are one way of looking at the market for theft and analyzing its costs. Another is as an example of the inefficiency of rent seeking. Both thieves and victims are competing for possession of the same objects--all of which, initially, belong to the victims. Expenditures by a thief either result in his getting the loot instead of some other thief or in his getting the loot instead of its owner keeping it. Defensive expenditures by the victims are also rent seeking--the function of a burglar alarm is to make sure that the property remains in the hands of its original owner.

What we have really been doing is showing the advantage of a system of secure property rights. If property rights are insecure, some individuals have an incentive to spend resources trying to get property transferred to them, while some have an incentive to spend resources keeping property from being transferred away from them. That is true whether the transfer is done by private theft or government taxation. Some of the inefficiencies we have just been discussing for the former case are known as excess burden in the latter; they were discussed at some length in Chapter 7. Not earning taxable income or not buying taxed goods are (costly) ways of defending yourself against taxation, just as installing a burglar alarm is a (costly) way of protecting against theft. Other inefficiencies on the political marketplace are classified as costs of lobbying. Making campaign donations to a candidate who promises to provide special benefits to you and your friends is an expenditure on transferring property in your direction almost precisely analogous to a burglar's expenditure on tools.

Yet another way of looking at theft--or rent seeking in general--is as a special case of the inefficiency of markets with externalities. Suppose auto firms spend $10,000 producing a car and in the process produce some air pollution. The result is inefficient--not because air pollution is evil but because it is not, like other costs, included in the firm's calculation of whether and at what price to produce. If the industry is competitive, everyone to whom a car is worth at least $10,000 gets one. If the air pollution does $5,000 worth of damage, then the real production cost is $15,000; anyone who values the car at more than $10,000 but less than $15,000 is getting something that is worth less to him than it costs to produce.

Theft is an extreme case of this: The external cost, imposed on the victim, is the entire value of what is stolen. The thief steals up to the point where the (marginal) value to him of what he steals is equal to the (marginal) cost to him of stealing it; since he ignores the cost to the victim, equilibrium occurs at a point where marginal cost to all concerned is much larger than marginal benefit.

We have now used economics to analyze the market for theft. In doing so, we took the system for preventing theft--police and punishments--as a given, one of the elements determining the cost to the thief of stealing. The next step is to use our analysis to say something about how that system should work: what should be illegal, what the nature and amount of the penalty should be, and how much we should be willing to spend on catching and convicting thieves. All of those can be viewed simply as issues of economic efficiency. While you may not believe that that is all they are, you should find the attempt to analyze them from that standpoint an instructive exercise.




In discussing the market for theft, I used the word "value" to describe both the value of what was stolen to the victim and its value to the criminal, without distinguishing between them. If what is stolen is money, gold, bearer bonds, or other liquid commodities--things that can be easily bought and sold at about the same price--that is a reasonable way of using the word. In other cases, it is not.

I have spent 20 hours searching art galleries to find a painting I particularly like and then bought it for $100. Replacing it will require a similar effort and a similar expenditure, so a thief who steals it injures me by considerably more than $100. The thief himself will be lucky to get $50 for it; even if he finds the right gallery and the right buyer--one who does not recognize the painting and does recognize its quality--he will get what the gallery pays for paintings, not what it gets for them.

In such a situation, the value to the thief of what he steals is much less than its value to the victim. That is why in many societies, including our own, there are well-established procedures by which thieves sell things back to their owners. Kidnappers are an extreme example of this. They steal something--a person--whose only value to them is what they can get by selling it back to (representatives of) its "owner."

This divergence between value to victim and value to thief suggests another way of looking at the inefficiency of theft. If you have something that is worth more to me than to you, I have no need to steal it; I can buy it from you. Goods that a thief is willing to steal but would not be willing to buy must be worth more to their present owner than to the potential thief. So the additional transfers that become possible as a result of theft are inefficient ones--transfers of a good to someone who values it less than its present owner. That is why I asserted, earlier, that the efficient level of theft was zero.

There are exceptions--what we may call "efficient crimes." Suppose, for example, that you are lost in the woods and starving. You come upon an empty, locked cabin. You break in, feed yourself, and use the telephone to summon help. Almost certainly, the value to you of using the cabin was greater than the cost you imposed on its owner; you will probably be glad to replace both his food and his lock. Your "crime" transferred a resource--temporary control of the cabin--to someone to whom it was worth more than its value to the initial owner. You could only do it by a crime, not by purchase, because the owner was not there to sell it to you.

This is one example of an efficient crime, a crime that is a net Marshall improvement. A less exotic example is speeding when you are in a hurry. We have speed limits, at least in part, because driving fast increases the chance of an accident. That is a cost but not an infinite one; there are times when getting somewhere quickly is sufficiently important to justify doing so at 80 miles per hour. One way of dealing with such situations is for the law to make it illegal to drive faster than 70 miles per hour except when there is an important reason to do so. The problem with such a law is that it may be difficult or impossible for a court to judge whether your reason was really good enough to justify an exception. An alternative solution is for the law to impose a penalty large enough so that only those who really have a good reason to drive faster will find it worth breaking the law and paying the penalty. Speeding is then always a crime--but if the punishment is correctly calculated, it is a crime that occurs when and only when it is efficient.

Seen in this way, a speeding law is a Pigouvian tax, like the emission fee discussed in Chapter 18. If air polluters must pay an emission fee equal to the damage done by the pollution, they will choose to pollute--and pay--only when the value of what is being produced is greater than the cost, including the cost of the pollution. Similarly, if driving fast imposes costs on other drivers, we can use speeding tickets to force motorists to take account of those costs in deciding how fast to drive.

The analysis so far suggests a simple rule for setting punishments: "The amount of the punishment should equal the damage done by the crime." That way, only efficient crimes will be committed--crimes for which the value to the criminal is greater than the amount of damage done.

One thing wrong with that rule is that criminals are not always caught. If a criminal faces only one chance in ten of being caught and convicted, he will discount the punishment accordingly in calculating the cost to himself of committing the crime. Roughly speaking, the punishment should then be ten times the damage done by the crime in order to assure that only efficient crimes, crimes for which the gain to the criminal is greater than the cost to the victim, occur. A more precise rule would have to allow for the criminal's attitude toward risk, as discussed in the optional section of Chapter 13; if the criminal is risk averse, one chance in ten of losing $100 may from his standpoint be more than equivalent to a certainty of losing $10.

This raises an interesting problem. The enforcement system can choose among many different combinations of probability and punishment; the same deterrence may be provided by a certainty of a $1,000 fine, a 50 percent probability of a $2,000 fine, a 10 percent chance of a $10,000 fine, or any of a variety of other alternatives, including some in which the punishment is not a fine but imprisonment or execution. How should it decide which to use?

The answer is the same as the answer to a similar problem many chapters back: the problem of choosing the mix of inputs for producing an output. Just as in Chapter 9, the first step is to pick a level of output and find the least costly way of producing it. In our present situation, the output is deterrence; it is produced by imposing a cost on the criminal. Pick a punishment--a fine of, say, $100 imposed with certainty on anyone who commits the crime. Consider all the combinations of (lower) probabilities and (higher) punishments that the criminal considers equivalent to a certainty of paying $100--and that will therefore have the same effect in discouraging him from committing the crime. For each combination, calculate the cost of catching that fraction of the criminals and imposing that punishment on each. Find the combination for which that cost is lowest. You now know the cost of the least-cost way of imposing an expected punishment equivalent to a $100 fine. Now pick another punishment--say a fine of $200 imposed with certainty--and repeat the calculation. When you are finished, you will have a total cost curve for deterrence and you will know, for any level of deterrence, what combination of probability and punishment should be used to produce it.

Two sorts of costs go into the calculations I have just described: enforcement costs and punishment costs. The nature of enforcement costs is fairly clear; they represent the expense of paying police, running the courts, and the like. But what are punishment costs?

Consider a fine. The convicted criminal pays $1,000, the court system collects $1,000. The money may be used to pay the expenses of operating the court system, given to victims of crime, or in some other way transferred, directly or indirectly, to someone other than the criminal. What the criminal loses someone else gains, so the net cost is about zero.

Suppose that instead of fining the criminal we execute him. The cost to him is his life; there is no corresponding gain to the rest of us. Suppose that, instead of fining or executing the criminal, we simply lock him up. The cost to him is his freedom. We get nothing--worse still, we must pay the cost of running the prison. The cost of the punishment--the cost to him plus the cost to us--is greater than the amount of the punishment.

As a rule, it is easier to collect small fines than large ones, since criminals are more likely to be able to pay them. So, as a general rule, punishment cost increases with the size of punishment. Enforcement cost, on the other hand, is larger the larger the percentage of criminals you are trying to catch. As we move from small punishments imposed with high probability to large punishments imposed with small probability, we are trading off a decrease in enforcement cost against an increase in punishment cost; somewhere between one extreme (catching 100 percent of the criminals and making them give back what they stole) and the other (catching only one criminal and boiling him in oil), there is likely to be an optimal combination.

Figure 20-3 shows the logic of the situation graphically. We are varying the probability of punishment from 0 to 1, while maintaining a constant level of expected punishment (equal to probability times punishment: pxF) and hence of deterrence; the lower the probability (p), the higher the punishment (F). The horizontal axis shows both probability and punishment; the vertical axis shows cost. The curve EC is enforcement cost. The larger the fraction of criminals we want to catch the more we must spend on police and courts, so EC rises as probability increases. The curve PC is punishment cost. The more severe the punishment the higher the cost of imposing it, so PC increases with amount of punishment (and thus decreases with probability of punishment). EC+PC is the total cost of catching and punishing criminals. Its minimum is at p*, so p* (and the corresponding punishment F*) represent the lowest cost combination of probability and punishment to produce a given expected punishment and hence a given level of deterrence. By repeating the calculation for every level of deterrence we could generate a total cost curve TC(d), showing the cost of producing any level of deterrence, d.

Calculating the least cost combination of probability and punishment to produce a given level of deterrence. pxF=100.

Our analysis has still not fully taken account of the costs of catching and punishing criminals. Consider a crime that does $10 worth of damage each time it is committed. If there is no punishment at all, 100 crimes per year will be committed, so the total damage will be $1,000/year; the net cost, including the benefits to the criminals and assuming the victims spend nothing on protecting themselves, will be between zero and $1,000.

The police force would like to impose an expected punishment of $10 on the criminals, as suggested by our previous discussion. The least expensive way to do this turns out to be catching one tenth of the criminals and fining them $100 each. Unfortunately, the cost of doing so is $2,000/year. The "efficient" level of punishment involves spending $2,000 to eliminate less than $1,000 of net damage! Obviously, that is the wrong answer; in the situation as described, the least bad solution may be to put up with 100 crimes a year.

As this suggests, the full analysis of what should be a crime, what percentage of the criminals should be caught, and what should be done to them is moderately complicated. The answer depends on the supply curve for offenses, on the damage done by the crime, and on the cost functions for producing apprehensions, convictions, and punishment. In principle, we now know how to solve the problem--just as, in principle, we know how to calculate how much a firm should and will produce, or what prices and quantities will be in a competitive equilibrium.

How do you calculate the efficient solution for the problem of preventing crime? Start by writing a cost function that includes costs and benefits to everyone affected. Next assume criminals and victims will choose values for the variables they control--amount of crime and amount of private defense against crime--that maximize their own welfare. Finally, find values for the remaining variables--percentage of criminals convicted, nature and amount of punishment--that maximize net benefits.



When we talk about law enforcement, we usually mean law enforcement by police officers. In fact, much of law enforcement is private. If someone breaks your arm, you call the police; but if he breaks a window or a contract, you call a lawyer. In the one case, law is enforced by government employees who gather evidence, present it to the court, collect the fine, and run the prison or close the switch on the electric chair. In the other case, law is enforced by a private individual, working for pay or a share of the settlement; he is responsible for gathering evidence and presenting it to the court, and he and his employer, the injured party, receive the "fine" paid by the convicted offender.

In our system, the division between public and private enforcement roughly corresponds to the division between criminal and civil law. Criminal law involves police, district attorneys, and sentences for criminals; civil law involves private detectives, private attorneys, and damages paid by defendants to plaintiffs. The form is in many ways different, but the substance is similar. In both cases it is alleged that someone has done something he should not have, and in both something unpleasant happens to the convicted defendant--whether we call it punishment or paying damages.

This raises some interesting questions about our system. Is there something natural about the present division into public and private enforcement? What are the advantages and disadvantages of the two systems? Could we have a system in which all law enforcement was public, so that a businessman who failed to deliver goods on time would be arrested, indicted, and jailed? Could we have a system in which all enforcement was private, so that a murderer would be sued for damages by the heirs of his victim?

Whether our present system is in some sense natural or efficient is a subject of dispute among economists involved in the economic analysis of law; my own belief is that it is not. What is clear is that different divisions between private and public enforcement are possible and have existed in other societies and at other times. They include some in which enforcement was entirely private; killing someone resulted in a lawsuit instead of an arrest. Whether or not we have the correct mix of private and public enforcement, it is clear that both systems have advantages and disadvantages. One of the inherent disadvantages of public enforcement is illustrated by the following immoral tale.

You are a police officer. You have got the goods on me. You have collected sufficient evidence to arrest and convict me; the resulting punishment would be equivalent, to me, to a $20,000 fine. Perhaps the punishment is a $20,000 fine; perhaps it is a period of imprisonment that I would pay $20,000 to avoid. For the purposes of the story, we will assume the former.

Arresting me will improve your professional reputation, slightly increasing your chances of future promotion. That is worth $5,000 to you in increased future income. Seen from the viewpoint of Dragnet, the rest of the story is clear; you arrest me and I am convicted. Seen from the viewpoint of this book, the result is equally clear. You have something--the collected evidence against me--that is worth $5,000 to you and $20,000 to me. Somewhere between $5,000 and $20,000, there ought to exist a transaction in our mutual benefit. I pay you $10,000, and you burn the evidence.

So far as you and I are concerned, that is an eminently satisfactory outcome, but it is not a very effective way of enforcing the law. In this respect, the public enforcement system is not incentive compatible. The system requires you to do something--arrest me--in order for it to work, and the system makes it in your interest to do something else. The system, of course, can and will try to control the problem--for example, by punishing police officers who are caught accepting bribes. But the fact that it must devote some of its limited resources to catching police officers instead of catching criminals is itself a defect.

Another way to solve the problem is to pay you, not a wage, but the value of the fines collected from the criminals you convict. Under such a system, you lose $20,000 when you burn the evidence, so $20,000 is the lowest bribe you will accept. Since $20,000 is also the cost to me of being convicted, there is little point in my offering you that much to let me off--save perhaps as a way of saving the time and expense of standing trial. If I do bribe you, no damage has been done; I have still paid $20,000 and you have still received it. We have merely eliminated the middleman.

This may sound like an odd and corrupt system, but it is the way in which civil law is presently enforced. The enforcing is done by a lawyer, acting as the agent of the victim; the fine is paid by the defendant to the victim. What we call bribery in criminal law is called an out-of-court settlement in civil law. The only addition to my scheme needed in order to make it correspond exactly to ordinary civil law is to make the claim against the criminal start out being the property of his victim; the police officer--who in this system is a private entrepreneur rather than a government employee--buys the claim from the victim before hunting down the criminal.

Elements of such a system for enforcing criminal law existed in the U.S. in the last century, as shown by the "Wanted Dead or Alive: $200 Reward" posters familiar in films and books about the Wild West. The policemen of that system were called bounty hunters. A complete system of private enforcement existed in Iceland in the early Middle Ages. Not only was killing treated as a civil offense, but the enforcement of court verdicts, including the job of hunting down convicted defendants who refused to pay and were consequently declared outlaws, was left to the plaintiffs and their friends. Odd as it may seem, the system appears to have worked fairly well; the society of which it was a part was one of the most interesting and in some ways one of the most attractive then existing. It was the source of the original sagas--historical novels and histories written in the thirteenth and fourteenth centuries and in many cases still in print today, in English translations.

Private enforcement has some advantages over public enforcement. It also has some problems. One is that many criminals are judgment-proof: They lack the assets necessary to pay any large fine. A public enforcement system can punish such criminals by imprisoning (or, in extreme cases, executing) them, but it is not immediately obvious how a private enforcer can make a profit that way. One cannot get blood from a turnip, and while a pound of flesh may add drama to a Shakespearean play, its market value is near zero. If a private enforcer cannot make money out of catching criminals, he has no incentive to do so, just as there is little incentive in our civil system to sue someone, however guilty, if he obviously cannot pay.

In analyzing the choice between private and public enforcement, as in discussing the problem of optimal punishment earlier, we get into complications that cannot be adequately dealt with in this book; so I will leave unresolved the question of whether our system of enforcement should be more private or less so. Some further discussions, including two of my articles, are listed at the end of the chapter .

Economics Joke #4: Incentive Incompatibility.

Jose robbed a bank and fled south across the Rio Grande, with the Texas Rangers in hot pursuit. They caught up with him in a small Mexican town; since Jose knew no English and none of them spoke Spanish, they found a local resident willing to act as translator, and began their questioning.

"Where did you hide the money?"

"The Gringos want to know where you hid the money."

"Tell the Gringos I will never tell them."

"Jose says he will never tell you."

The rangers all cock their pistols and point them at Jose.

"Tell him that if he does not tell us where he hid the money, we will shoot him."

"The Gringos say that if you do not tell them, they will shoot you."

Jose begins to shake with fear.

"Tell the Gringos that I hid the money by the bridge over the river."

"Jose says that he is not afraid to die."


The economic analysis of accidents starts with the observation that they are not entirely accidental. I do not choose to run my automobile into a pedestrian, but I do choose what kind of car I drive, how often and at what speed I drive it, and how often to have my brakes checked. These decisions and many more affect the probability that I will be in an accident and thus the cost my driving imposes on other people. It then seems natural to ask what set of legal rules will lead me to make such decisions in the most nearly efficient manner.

The simplest approach to generating efficient behavior is direct regulation. Let the law state how cars must be built, how many miles people may drive and at what speed, how often their brakes must be checked. This solution runs into problems that we have already discussed, in the context of monopolies, public goods and externalities. In order to set efficient values for all of the variables, the legislature would require detailed information about individual tastes and abilities that it has no way of getting; if it had the information, using it to calculate optimal behavior would involve daunting mathematical problems. Even if the legislature could calculate and enforce optimal behavior, there is no obvious reason why it would be in the interest of the legislators to do so.

Similar problems arise if we try to control accidents by Pigouvian taxes on the behavior that causes them. The probability that I will be involved in an accident depends on many choices, only some of which are observable. If by driving an extra mile I impose an expected cost of five cents on potential accident victims, then a tax of five cents a mile will induce me to do the efficient amount of driving. But what tax will prevent me from paying a more than optimal amount of attention to the radio and a less than optimal amount to the road?

The solution to this problem is to charge by results: If I cause an accident I must pay the cost. Externalities are then internalized; I have an incentive to engage in an efficient level of accident provision on every margin. In our legal system, such costs are imposed mainly through civil suits for damages.

This produces new problems. Driving becomes a lottery with large negative prizes. If drivers are risk averse they have an incentive to insure themselves--and, by doing so, reduce their incentive to take precautions. Many drivers will be judgment-proof, unable to pay the cost of a major accident. That can be solved by requiring drivers to be insured, but again with negative effects on incentives.

There is another and deeper problem. As Coase pointed out, the Pigouvian approach contains a fundamental error; it treats external costs as if they were the result of only one party's action. The probability that I will run you down depends on your decisions as well as mine, on how carefully you cross the street as well as on how fast I drive. Ideally both of us should take all cost-justified precautions. But if the driver is responsible for all costs of the accident, the pedestrian has no incentive to take precautions.

There are at least three solutions to this problem. One is a rule of negligence. The driver is liable if and only if he is negligent, with negligence defined as failure to take all cost-justified precautions. Under that rule, drivers find it in their interest to take the efficient level of precautions; pedestrians, knowing that drivers will not be negligent and thus not be liable, find it in their interest to take an efficient level of precautions as well.

A second solution is a rule of strict liability with a defense of contributory negligence. A pedestrian who has taken less than the efficient level of precautions cannot collect damages from the driver. This rule again leads to an efficient level of precaution by both parties.

Both of these suffer from the same problem as direct regulation. In order to apply either negligence or contributory negligence, the court must be able to calculate the efficient level of precaution and observe whether it is taken. They are an improvement only in restricting the problem to cases where an accident actually occurs.

Even if the court can judge whether the driver was negligent in how he drove, it can hardly judge whether he was negligent in how much he drove--whether his marginal trip was worth taking, given the expected accident costs it produced. Under a negligence rule, drivers drive too much since, having taken the efficient level of precaution, they are no longer liable for damages. Under a rule of strict liability, with or without contributory negligence, there is an efficient amount of driving but an inefficiently large amount of walking, since pedestrians are reimbursed by drivers for the cost of accidents.

The solution to this problem carries us outside of the context of civil damages. If the driver pays damages but the pedestrian does not receive them--more generally, if each party to the accident must separately pay its full cost--then each has the efficient incentive to avoid the accident. The damage award has been converted into a fine.

This solution generates new problems. If the injured party receives no damages he has no incentive to sue, so the accident never gets reported. In converting damages into fines we have gone from a private to a public system of law, and must provide some public mechanism to report damages and institute cases. In doing so, we encounter precisely the same problems that I discussed earlier as arguments for converting public enforcement into private enforcement.

The essential problem in accident law, as in more direct forms of government regulation, is that we have no bureaucrat-gods. If only there were an institution, whether regulatory agency or court, that knew everything and wished only the general good, the production of efficient outcomes would be fairly simple. As it is, we must choose among a bewildering variety of imperfect solutions, private and public, criminal and civil. The economic analysis of law helps us to understand the problem but it does not, at least so far, give us any clear answer.

In fact, I have not given final answers to many questions in this chapter, nor have I solved many problems. What I hope I have done is to convince you that economic analysis can be used to evaluate such fundamental issues as what the laws should be, what the penalties should be for breaking them, and how those penalties should be enforced. The economic analysis of law is an important part of what some of us like to call economic imperialism--the use of economics to analyze what have traditionally been considered "noneconomic" questions--and, as you may have guessed, one that I have found particularly interesting.



1. Suppose Nicholas Van Rijn applied his talents to the problem of preventing airplane hijacking. What might he suggest?

2. There is a way to make theft forever impossible--perhaps by adding a "guilt drug" to the water supply that would make anyone who steals anything feel intolerably guilty. Would all, some, or none of those who are presently thieves be in favor of doing this? Discuss.

3. In Chapter 18, I discussed problems associated with software piracy--the unauthorized copying of computer programs. As the name suggests, this can be regarded as a form of theft; indeed, I once shocked one of my colleagues by suggesting that his possession of disks full of pirated software made him the moral equivalent of a burglar. Suppose there is some way of making such software piracy impossible. Should none, some, or all of those who presently use pirated software be in favor of the change? Discuss.

4. In this chapter, I have argued that certain things should be illegal on grounds of economic efficiency. If this is our criterion, what things that are presently illegal should not be? Should they be illegal on other grounds? If so, why? Discuss.

5. Throughout the chapter, I have ignored the possibility that some people might abstain from stealing because they thought it was immoral. How could we include that possibility in the analysis? Would it change any of the results? Discuss.

6. a. You see a highway sign that says "$500 Fine for Littering." What is the economic rationale for having a fine on littering and why is it so high?

b. On the same road, the fine for speeding is only $100. Does that mean that the government is more concerned about littering than about speeding? Discuss.

c. In what senses might a fine for speeding be too high? Discuss.

7. "We can always lower the cost of our criminal justice system by catching half as many criminals and punishing them twice as harshly; the system will be just as effective as before at deterring crime and we will not need to hire as many police officers." Discuss.

8. Some legal scholars object to the economic analysis of law on the grounds that laws should be just, not efficient. This may be seen either as the claim that consequences do not matter (Fiat justitia, ruat caelum--let there be justice though the sky fall) or that consequences do matter, but economic efficiency is the wrong way of judging them. Discuss.



My analysis of private enforcement is in "Efficient Institutions for the Private Enforcement of Law," Journal of Legal Studies (June, 1984), which also contains references to earlier work on the subject by others, not all of whom agree with me. The Machinery of Freedom, cited in the previous chapter, contains a nontechnical discussion of how a fully private system of courts, police, and laws might work.

My earlier article, "Private Creation and Enforcement of Law--A Historical Case," Journal of Legal Studies (March, 1979), describes the working of the Icelandic system, as does Feud in the Icelandic Saga (Berkeley: University of California Press, 1982) by Jesse Byock (a historian, not an economist). My "Reflections on Optimal Punishment Or: Should the Rich Pay Higher Fines?" in Richard Zerbe (ed.), Research in Law and Economics, Vol. 3 (1981) contains a detailed analysis of optimal punishment. My essay "Economic Analysis of Law" in The New Palgrave: A Dictionary of Economic Theory and Doctrine, John Eatwell, Murray Milgate and Peter Newman, eds. (Macmillan, 1987) gives a general overview of the subject and an extensive list of references.

Other works of interest include: Norval Morris and Gordon Hawkins, The Honest Politician's Guide to Crime Control (Boston: Little, Brown & Co., 1977); Gordon Tullock, The Logic of the Law (New York: Basic Books, 1971) and Richard Posner, Economic Analysis of Law 3rd Edn. (Boston: Little, Brown, 1986). Posner, who is one of the leading writers on the economic analysis of law (and a federal judge), argues that the common law tends, for a variety of reasons, to be economically efficient. If he is right, then efficiency may be useful for explaining what the law is, whether or not it is useful for deciding what it ought to be.

The Last Testament of Lucky Luciano, by Martin A. Gosch and Richard Hammer (Boston: Little, Brown: 1974), claims to be based on information given to Gosch by Luciano on condition that it not be used until ten years after his death. It is an interesting, and on the whole plausible, account of the workings of organized crime from the viewpoint of one of its leading members.

"Fact, fancy, and organized crime", by Peter Reuter and Jonathan B. Rubinstein, The Public Interest 53 (Fall 1978) pp. 45-67. This article provides evidence and arguments that support my view of organized crime, including the results of the study of bookmaking and numbers mentioned in this chapter.

Next Chapter

Table of Contents