Econ 522 Economics of Law Dan Quint Fall 2016 Lecture 22

Logistics HW4 due next week Final exam Fri Dec 23, 10:05 a.m., in this room Fully cumulative – generally more weight on second half of semester, but everything is fair game! Online course evaluations are open until Friday Office hours tomorrow: 3-4 1 1

Back to criminal law

So far… If we assume rational criminals… …then crimes whose benefit is less than the expected punishment (probability X severity) are deterred Deterrence might cost money (hire more police, etc.)… …but it can also save money, because fewer crimes could mean fewer people you have to punish 3 3

+ So far… X X We’ve been focusing on optimal enforcement How much should we invest in catching criminals? How should we punish them when we catch them? Very little on determining guilt or innocence Get the facts, decide how likely it is they committed the crime Based on relative costs of freeing the guilty and punishing the innocent, there’s some amount of certainty above which we punish In theory, we just want to minimize But how do we do that? Probability of wrongful exoneration Social cost of wrongful exoneration + Probability of wrongful conviction Social cost of wrongful conviction X X 4 4

Determining guilt or innocence http://www.youtube.com/watch?v=zrzMhU_4m-g starting at 1:08 ending at 3:07 5 5

Peter Leeson, “Ordeals” (2012, Journal of Law and Economics) First in a series of papers on “the law and economics of superstition” “For 400 years the most sophisticated persons in Europe decided difficult criminal cases by asking the defendant to thrust his arm into a cauldron of boiling water and fish out a ring. If his arm was unharmed, he was exonerated. If not, he was convicted. Alternatively, a priest dunked the defendant in a pool. Sinking proved his innocence; floating proved his guilt. People called these trials ordeals. No one alive today believes ordeals were a good way to decide defendants’ guilt. But maybe they should.” 6 6

Ordeals Ordeals were only used when there was uncertainty about guilt or innocence Examples: hot water ordeal hot iron ordeal cold water ordeal Sounds crazy, but Leeson’s point: ordeals may have actually done a pretty good job of ascertaining guilt/innocence! hot water ordeal: fish a stone or a ring out of a pot of boiling water; if your hand gets burned, you’re guilty hot iron ordeal: carry a piece of burning iron a certain distance cold water ordeal: tied up, thrown in water (and pulled out); if you float, you’re guilty 7 7

Why would ordeals work? iudicium Dei – Medieval belief that God would help the innocent survive the ordeal, but not the guilty If people believe this, then… guilty won’t want to go through with the ordeal, will instead confess (confessing leads to a lesser punishment than failing the ordeal, plus you don’t burn your hand) the innocent agree to go through the ordeal, expecting to be saved by a miracle So administering priest knows if someone agrees to take the ordeal, he’s innocent… …and rigs the ordeal so he’ll pass 8 8

The keys to this working Obviously, people have to believe that God will spare the innocent, judge the guilty Ceremony reinforced this by linking the belief to other religious beliefs Priests might have to let someone fail an ordeal once in a while, to keep people believing (Again, this was only done in cases where normal evidence was lacking, so unlikely to be “contradicted”) And, priests must have a way to rig the ordeals Leeson gives examples of how the ordeals were designed to make this easy: priests were alone before and after the ordeal, spectators couldn’t be too close, priests had to judge whether the person had passed or not, etc. 9 9

“Evidence” to support Leeson’s view Historically, most people who underwent ordeals passed Data from 13th century Hungary: 130 out of 208 passed England, 1194-1219: of 19 for whom outcome was recorded, 17 passed And this seems to have been by design Other historians: “the ordeal of hot iron was so arranged as to give the accused a considerable chance of escape.” Others: “the average lean male has an 80% chance of sinking in water, compared to only a 40% chance for the average lean woman.” England, 1194-1208 84 men went through ordeals – 79 were given cold water ordeal 7 women went through ordeals – all were given hot iron ordeal 10 10

“Evidence” to support Leeson’s view Ordeals were only used on believers “If the defendant was Christian, he was tried by ordeal. If he was Jewish, he was tried by compurgation instead.” Once Church rejected legitimacy of ordeals, they disappeared entirely 11 11

Leeson’s conclusion “Though rooted in superstition, judicial ordeals weren’t irrational. Expecting to emerge from ordeals unscathed and exonerated, innocent persons found it cheaper to undergo ordeals than to decline them. Expecting to emerge… boiled, burned, or wet and naked and condemned, guilty persons found it cheaper to decline ordeals than to undergo them. [Priests] knew that only innocent persons would want to undergo ordeals… [and] exonerated probands whenever they could. Medieval judicial ordeals achieved what they sought: they accurately assigned guilt and innocence where traditional means couldn’t.” Like I said, this was paper 1 out of 3 on law and econ of superstition The second paper is on “judicial battle” – before 1179 in England, property disputes would be settled by having the two sides hire knights to fight each other The third paper is about gypsies 12 12

Deterrence

Do harsher punishments deter crime? Hard to answer, because hard to separate two effects Deterrence When punishment gets more severe, crime rates may drop because criminals are afraid of being caught Incapacitation When punishment gets more severe, crime rates may drop because more criminals are already in jail Cooter and Ulen discuss several empirical attempts to measure the extent to which punishment deters crime. This is hard, because it’s often difficult to separate two separate effects: deterrence – when punishment gets more severe, crime rates may go down because people are more afraid of being caught incapacitation – when punishment gets more severe, say, through longer prison terms, so crime rates may go down just because more criminals are already in jail The early literature didn’t deal so directly with this problem, but did find that higher conviction rates and harsher punishments were associated with lower crime rates. Two studies, rather than looking at high-level crime rates, studied individual people who were likely to commit crimes: convicted criminals being released from prison Within this population, there was still a deterrence effect: Those with a high chance of being convicted again were arrested less in the months following release.

Do harsher punishments deter crime? Hard to answer, because hard to separate two effects Deterrence When punishment gets more severe, crime rates may drop because criminals are afraid of being caught Incapacitation When punishment gets more severe, crime rates may drop because more criminals are already in jail Kessler and Levitt: natural experiment Voters in California in 1982 passed ballot initiative adding 5 years per prior conviction to sentence for certain crimes Found immediate drop of 4% in crimes eligible for enhanced sentences One paper (by Dan Kessler and Steve Levitt) used a “natural experiment” to separate the deterrent effect from the incapacitation effect In 1982, voters in California passed a ballot initiative which added 5 years to sentences for certain serious crimes for each prior conviction by the criminal Any immediate change in crime rates should be due to deterrence, since the number of criminals in prison would only respond gradually They found an almost immediate 4% drop in the crimes that were eligible for these “sentence enhancements” So there did seem to be a clear deterrent effect

Probability versus severity Empirically, crime levels more sensitive to probability of being caught, than to severity of punishment Might be that criminals discount future a lot, don’t care as much about last few years of a long prison sentence… …or, total cost of punishment may be more than “apparent” sentence Punishment = time in jail… …plus other costs – time spend in jail awaiting trial, money spent on a lawyer, stigma of being a convicted criminal… …which may not depend on length of sentence So 20% X (1 year in jail + C) > 10% X (2 years in jail + C) not because 20% of 1 year is worse than 10% of 2 years but because 20% of C is more than 10% of C Means Becker’s idea – tiny probability, very severe punishments – may not work in real life There is some empirical evidence that harsher expected punishments do have a deterrent effect But the results are interesting: on the margin, crime levels seem to respond much more to changes in the probability of getting caught, than to changes in the severity of punishment once you’ve been caught That is, doubling the chance of being caught has more of an effect on crime rates than doubling the fine or the prison term that convicted offenders will face. Friedman gives a simple argument for why this isn’t surprising: the total private cost of being punished is actually different from the “apparent” sentence If someone is convicted and sentenced to 1 year in jail, they already spent some time in jail awaiting trial, spent money on a lawyer, and will have to live with the stigma of being a convicted criminal These additional penalties are similar regardless of the length of the prison sentence So 20% X (1 year + C) > 10% X (2 years + C) not because 20% of 1 year is worse than 10% of two years, but because 20% of C is greater than 10% of C. But the fact that people respond more to the probability of getting caught than to the sentence seems to suggest that Becker’s suggestion – very harsh punishments with very low probabilities – wouldn’t work well in practice (Also working against Becker is the fact that people might object to punishments that are way out of proportion with offenses If the chance of being caught jaywalking is very small, we might have to execute jaywalkers in order to deter jaywalking But do we really want to execute jaywalkers? And would juries really convict jaywalkers if the penalty was death?)

Marginal deterrence Armed robbery vs. armed robbery plus murder You break in to rob an isolated house, carrying a gun Someone wakes up and confronts you; what do you do? Punishment for murder is very severe If punishment for armed robbery is not so severe, you might leave them alive If punishment for armed robbery is very severe, you might be better off killing them One simplification we’ve been making is that we’re only considering one crime at a time. But if different crimes are substitutes for each other, increasing the punishment on one may push offenders into committing the other crime instead. Take the example of armed robbery, and armed robbery plus murder You break into an isolated house, carrying a gun, intending to steal stuff Someone hears you and wakes up What do you do? If the punishment for armed robbery is much less severe than the punishment for murder, you might choose to run You take your chances on the owner calling the police and being able to identify you And you know that at least the punishment won’t be that drastic if you’re caught If the punishment for armed robbery is very high, you might be better off shooting the guy This reduces your chance of getting caught And you know that the punishment won’t be that different if you are caught

Marginal deterrence Armed robbery vs. armed robbery plus murder You break in to rob an isolated house, carrying a gun Someone wakes up and confronts you; what do you do? Punishment for murder is very severe If punishment for armed robbery is not so severe, you might leave them alive If punishment for armed robbery is very severe, you might be better off killing them Same argument against “three strikes” laws “As good be hang’d for an old sheep as a young lamb.” - old English proverb This argument has been used to argue against “three strikes” rules, which mandate long (20 years or more) prison terms for third convictions on certain types of crimes. If someone is about to get caught for a relatively minor offense, facing a very severe punishment might make them more willing to kill someone to avoid being caught. (The logic is captured in the old proverb, “As good be hang’d for an old sheep as a young lamb.” In England, the penalty for either was death; since an old sheep has more meat and wool than a young lamb, since the penalty is the same, you’re better off stealing the more valuable animal.) (This is a useful way to think about the example where someone drove over a baby a second time. Implicitly, the “punishment” for killing someone was less than the punishment for injuring someone and having to pay for their medical care – so the marginal deterrence was actually negative!)

Punishment

Punishment In U.S., most crimes punished by imprisonment Imprisonment has several effects: Deterrence Punishment Opportunity for rehabilitation Incapacitation When is incapacitation effective? When supply of criminals is inelastic (When there isn’t someone else waiting to take criminal’s place) And when it changes number of crimes a person will commit, rather than just delaying them Incapacitation is only effective under certain situations First, incapacitation only matters if an arrested criminal won’t be immediately replaced by someone else If you arrest the head of a drug gang, his top lieutenant might take over, and crime might go on apace Cooter and Ulen put it this way: incapacitation is effective at reducing crime when the supply of criminals is inelastic Second, incapacitation only matters if it changes the number of crimes a person will commit, rather than just delaying them until he gets out of prison If punishment for a third offense is very severe, most criminals might choose to only commit crimes until they’ve been caught twice A longer sentence for the first offense may delay the second round of crimes, but not eliminate it On the other hand, crimes that are caused by impulsive youth, and whose motivation fades with age, would be lowered by incapacitation.

Punishment Fines are efficient No social cost But, greater threat of abuse, since state makes money Friedman: “In a world of efficient punishments, somebody gets most of what the convicted defendant loses. It is in that somebody’s interest to convict defendants, whether or not they are guilty.” Other punishments tend to be inefficient Direct costs of holding someone in maximum-security prison estimated at $40,000/year In some states, prisoners do useful work Attica State Prison (NY) had metal shop Minnesota firm employs inmates as computer programmers Medium-security prisons in Illinois make marching band uniforms But there are legal limitations on using prisoners as labor

Discretion in sentencing 1980-1990: move from judicial discretion to mandatory sentencing most state and federal courts sentence mostly pinned down by crime and offender’s history Recent move in opposite direction MI, LA recently brought back discretionary sentencing MS brought back discretionary parole for nonviolent first-time offenders Various sentencing reforms in 18 other states Between 1980 and 1990, most state and federal courts moved from giving the judge wide discretion in sentencing toward mandatory sentencing In many cases, a combination of the seriousness of the crime and the offender’s past history pin down the exact punishment. Much of the increase in the prison population is traced to mandatory sentencing in drug cases. In recent years, due partly to overcrowded prisons and rising costs, there’s been some blowback. Michigan and Louisiana recently moved from mandatory sentencing back to discretionary sentencing Mississippi, which abolished discretionary parole in 1995, brought it back for nonviolent first-time offenders Eighteen other states have passed some sort of sentencing reforms.

Fines Western Europe: many crimes punished by fines Textbook cites a study from 1977 examining certain crimes: 56% of selected defendants in England/Wales, 77% in Germany were punished only by a fine U.S. federal court: 5% of defendants punished only by a fine In U.S., criminal fines are in dollars; in Europe, “day fines” Punishment = fixed number of days of salary So, rich pay bigger fines than poor

For example… Guy caught driving his Ferrari 85 mph in a 50 through a village in Switzerland Repeat offender Swiss speeding tickets proportional to wealth, driver was worth $23 million, so speeding ticket was $290,000

Should the rich pay bigger fines than the poor? Some crimes have monetary benefits… Stealing $100 has same monetary benefit for rich or poor So penalty with same monetary equivalent – say, $1,000 – should have same deterrent effect …some have nonmonetary benefits Punching someone in the face in a bar might have same utility benefit for rich or poor Since rich have lower marginal utility of money, it would take a larger fine to have same deterrent effect Friedman describes a great back-and-forth between him and some of his students, on the question of whether the rich should face higher fines than the poor for the same offense. Friedman argued that fines should be the same for everyone A particular offense imposes some amount of external damage Setting punishment equal to damage leads to efficient choices, by both rich and poor Friedman’s students, though, pointed out that there are two different types of crimes: those with monetary benefits for the offender, and those with nonmonetary benefits. First, consider stealing $100 This has the same monetary benefit to someone rich or someone poor So a penalty with the same monetary equivalent – say, a $1000 fine – would have the same deterrent effect on the rich or the poor. But second, consider the joy of punching someone in the face who looked at you funny in a bar Suppose this has the same utility benefit for both the rich and the poor Because the rich have a lower marginal utility of money, it would take a larger fine to have the same deterrent effect

Should the rich pay bigger fines than the poor? Some crimes have monetary benefits… Stealing $100 has same monetary benefit for rich or poor So penalty with same monetary equivalent – say, $1,000 – should have same deterrent effect …some have nonmonetary benefits Punching someone in the face in a bar might have same utility benefit for rich or poor Since rich have lower marginal utility of money, it would take a larger fine to have same deterrent effect But with costly enforcement, goal isn’t to deter all crimes Some examples: optimal to deter “most” crimes by both rich and poor, which requires higher fines for rich people Some examples: optimal to deter poor peoples’ crimes, not bother deterring crimes by the rich! But in fact, things are more complicated than that Remember that with costly enforcement, the goal is not to deter all crimes, but to deter most crimes – that is, deter crimes until it stops being worth it to deter the next one Friedman apparently had his students work through some examples They found that sometimes, it’s optimal to set fines to deter “most” poor peoples’ crimes and “most” rich peoples’ crimes – which involves setting higher fines for rich people But if deterring the rich is too costly, it’s sometimes optimal to deter poor peoples’ crimes and accept that the rich will do what they want! (Friedman gives two more complications, which could favor either higher or lower punishments for rich people.)

Should the rich pay bigger fines than the poor? Society may have other goals besides efficiency Might place high value on law treating everyone the same… …even if we have to sacrifice some efficiency to achieve this Example: choice of a fine or jail time Tend to put low dollar value on time in jail – might be sentenced to a $5,000 fine or a year in jail Most people who can afford the fine will choose to pay; those who can’t, will go to jail So rich pay a small-ish fine they can easily afford, and poor go to jail This also brings us to another important point: Sometimes, society may have other goals in mind besides efficiency We’ve been using efficiency as our proxy for goodness throughout this course But it might be that in criminal cases, society puts a high value on treating everyone the same, even if this doesn’t lead to the most efficient outcomes. People might not object too heavily if rich people were charged higher fines than poor people for similar crimes People would probably object very strenuously if poor people were given more severe punishments than rich people for similar offenses That is, society might put a high value on “equality before the law,” and be willing to sacrifice some efficiency to get it. One example of this: certain crimes are often punished by a choice between a fine and jail time These tend to put a fairly low dollar equivalent to time in jail – for instance, you might be sentenced to a $5,000 fine or a year in jail The disutility of a lost year is pretty large – the likely result of this is that everyone who can afford the fine will pay it, and everyone who can’t, will go to jail But this means that rich offenders pay a small fine they can easily afford, and poor offenders spend a year in jail, which may seem objectionable. (In fact, since fines have a much smaller social cost than imprisonment, it may be preferable to push people toward paying fines rather than jail time, so a rule like this may make sense. But it still may seem objectionable to some people.)

Should the rich pay bigger fines than the poor? Society may have other goals besides efficiency Might place high value on law treating everyone the same… …even if we have to sacrifice some efficiency to achieve this Example: choice of a fine or jail time Tend to put low dollar value on time in jail – might be sentenced to a $5,000 fine or a year in jail Most people who can afford the fine will choose to pay; those who can’t, will go to jail So rich pay a small-ish fine they can easily afford, and poor go to jail John Lott: equal prison terms for rich/poor may make sense The rich value their time more than poor… …but the rich have better lawyers, may be less likely to be convicted (Putting aside questions of equality, John Lott argued that equal punishments for the same crime might be appropriate when the punishment is imprisonment: Most crimes are punished by imprisonment The rich value their time more highly than the poor, so imprisonment may be a more severe punishment to the rich than to the poor This would suggest the rich could get shorter jail sentences and face the same deterrent But on the other hand, the rich, on average, have better lawyers So they get convicted with lower probability The two effects might balance out, so the same punishments might be appropriate.)

Stigma Stigma of having been convicted of a crime You’re a corporate treasurer, and get caught embezzling One consequence: you go to jail for a year Another: when you get out, you can’t get another job as a treasurer Punishment = jail time + stigma We said earlier that fines have no social cost, and most other punishments – jail, execution – have significant social costs Friedman points out one type of punishment has negative social cost, that is, that actually creates value. This punishment is the social stigma of having been convicted Suppose you’re a pretty good corporate treasurer, and one day, you get caught embezzling money One consequence is you go to jail for a year Another consequence is that when you get out, you find it hard to get another job as a corporate treasurer Your punishment consisted of both jail time and a stigma, that is, other peoples’ knowledge of your crime. Friedman points out two empirical studies by John Lott One studied white-collar criminals, and found that the lost income following conviction was a significant fraction of the punishment Another looked at corporations charged with cheating their customers, and found that the drop in stock value was many times larger than whatever actual punishment was imposed.

Stigma Stigma of having been convicted of a crime You’re a corporate treasurer, and get caught embezzling One consequence: you go to jail for a year Another: when you get out, you can’t get another job as a treasurer Punishment = jail time + stigma Stigma as a punishment has negative social cost No wage at which firm would hire an embezzler as treasurer So getting hired by that firm would be inefficient But without the conviction, you might have gotten the job So knowledge that you’re an embezzler has value to society Friedman claims that stigma as a punishment actually has negative social cost To show this, he proposes the following experiment First, get caught embezzling and go to jail Next, go interview for jobs Tell people that you’re well-qualified to be a corporate treasurer, but you recognize they may be hesitant to hire a convicted embezzler, so you’re willing to work for much less than the going rate Friedman predicts that there’s no price at which most firms would knowingly take on an embezzler as treasurer That is, the cost to them of hiring you (at any wage) is greater than the value to you of having the job This means that you getting hired by that firm would be inefficient Without the conviction, you might have gotten the job; with the conviction, you don’t So society is better off with you being known to be an embezzler (Or, the information that you are the type of person who embezzles is valuable to society.)

Stigma Stigma as a punishment Suggests that maybe… Very efficient (negative social cost) when applied to a guilty person Very inefficient when applied to an innocent person Suggests that maybe… criminal cases, where conviction carries a social stigma, should have higher standard of proof than civil cases, where it doesn’t Friedman points out that stigma has one other funny characteristic: As a punishment, it’s efficient when applied to a guilty person, but very inefficient when applied to an innocent person That is, if someone who’s not an embezzler gets convicted of embezzlement, they won’t get any jobs in the future, but this is inefficient This is one argument for why criminal cases, which tend to come with a social stigma, should carry a higher standard of proof than civil cases, which don’t come with the same stigma.

Death penalty In 1972, U.S. Supreme Court found death penalty, as it was being practiced, unconstitutional Application was “capricious and discriminatory” Several states changed how it was being administered to comply In 1976, Supreme Court upheld some of the new laws Since 1976, average of 41 executions per year in the U.S. Texas and Oklahoma together account for half Nationwide, 3,000 prisoners currently on death row Since 1976, 304 inmates on death row were exonerated, many more pardoned or had sentences commuted by governors Does death penalty deter crime? Evidence mixed. In 2003, outgoing governor of Illinois converted sentences of all 167 Illinois death row inmates to life in prison Numerous studies on whether the death penalty actually deters crime, results have varied widely Some have found no deterrent effect; others have found a strong one All the results appear to be very sensitive to the exact formulation of the empirical model (The results literally range from no effect at all on one end, to one study finding that each incremental execution prevents 150 murders, with many studies finding a much lower, but positive, level of deterrence.) One problem complicating things is that, in situations where the death penalty will almost certainly be applied, judges and juries are sometimes less willing to convict someone. There’s a recent paper by Steven Durlauf (here at Wisconsin) and three other authors They point out that existing results on deterrence are very sensitive to the exact details of the empirical model used Given that, they argue not that the results are useless, but that model uncertainty – that is, uncertainty over which model is correct – should be explicitly incorporated into the model They propose a way of doing this, and find a modest, though not particularly strong, deterrent effect. Durlauf’s explanation of the main results: If you believe the death penalty deters crime, the data won’t convince you it doesn’t; if you believe it doesn’t, the data won’t convince you it will. The direct costs of capital punishment are currently quite high, due to a number of legal protections: capital cases are longer more jurors are rejected, so more jurors have to be made available initially the trial is often divided into two phases, one for guilt and one for sentencing and defendants are also entitled to automatic appeals holding someone on death row is more expensive than keeping them with the rest of the prison population In the U.S., it’s currently more expensive to execute someone than to hold them in prison for life

Fraction of convicted murderers sentenced to death in one study Death penalty and race One concern about death penalty in U.S.: way it’s applied is racially biased McCleskey v Kemp (U.S. Sup Ct 1987) even solid statistical evidence of racial disparity does not make death penalty unconstitutional Fraction of convicted murderers sentenced to death in one study White Victim Black Victim Overall 7.9% 22.9% 2.8% Black Defendant (Source of numbers: http://math.stackexchange.com/questions/83756/examples-of-simpsons-paradox in the comments, who claims it is adapted from Subsection 2.3.2 of A. Agresti (2002), Categorical Data Analysis, 2nd edition, Wiley, pp 48-51. It cies a 1991 study by Radelet and Pierce; I couldn’t find that study. A different study by the same authors at about the same time is published as Michael Radelet and Glenn Pierce, “Choosing Those Who Will Die: Race and the Death Penalty in Florida”, Florida Law Review vol 43, Jan 1991; it gives similar results, but doesn’t illustrate Simpson’s Paradox as cleanly, so I’m sticking with this one.) Studies that look only at race of the defendant are mixed – some find that black defendants are more likely to be sentenced to death, others find that there’s no difference Some have even found that white defendants are more likely to be sentenced to death But once you consider the race of the victim, a different pattern emerges Since blacks are more likely to kill blacks, and whites more likely to kill whites, these differences are easy to miss “Simpson’s Paradox” Quoting a recent New York Times article (http://www.nytimes.com/2008/04/29/us/29bar.html) The majority opinion in McCleskey was written by Justice Lewis F. Powell Jr. After he retired, his biographer asked Justice Powell whether, given the chance, he would change his vote in any case. “Yes,” Justice Powell said. “McCleskey v. Kemp.” 11.0% 11.3% 0.0% White Defendant source: see notes.

Other topics

“Victimless” crimes Many crimes don’t seem to directly harm anyone Cannibalism (when victim is already dead), organ sales But in a world where cannibalism is legal, the private benefit of murder is higher Might lead to more murders Same with organ sales Once human organs become valuable, tradable commodities, value of killing someone becomes higher May make sense to outlaw certain practices that do no harm themselves, but encourage other harms Friedman explains criminal laws against some “victimless” crimes – cannibalism (once someone is already dead), organ sales, etc. In a world where cannibalism is allowed, the private benefit of murder is higher; so murder might be more common Even more so with organ sales – once human organs become valuable commodities, the value of killing someone becomes much higher, and may outweigh the risk of punishment Thus, it may make sense to outlaw certain practices that do no harm in and of themselves, but encourage other harms. (One might be able to make a similar case for illegal drugs, although I’m not sure the case would be any stronger than with alcohol.)

Drugs Increasing expected punishment for dealing drugs will increase street price What happens next depends on elasticity of demand Casual users tend to have elastic demand If price goes up (or risk/difficulty in obtaining drugs goes up), demand drops a lot Addicts have very inelastic demand Price goes up, but demand stays about the same So expenditures go up Drug addicts who support their habits through crime, have to commit more crimes “Ideal” policy might be to raise price for non-addicts without raising price for addicts Cooter and Ulen discuss a bit the efforts to control illegal drug use They point out that increasing the expected punishment for selling drugs will increase its price What happens then depends on the elasticity of demand. Casual (occasional) drug users tend to have fairly elastic demand When the price of drugs (or the risk or difficulty in obtaining them) goes up, their demand goes down significantly So punishing drug sales has the desired effect. Addicts, on the other hand, tend to have very inelastic demand Since they’re physically dependent on the drug, they still want to consume nearly the same amount So price goes up, but demand stays nearly the same What changes is expenditures – addicts must pay more for drugs For those who support their habits via crime, this can lead to more crime. Cooter and Ulen point out that the “ideal” policy might be to raise the price to non-addicts without raising the price to addicts Several programs in Europe to just that Addicts can register with the government and receive drugs cheaply Those who do not register face higher street prices Not sure I see that happening here so much Cooter and Ulen also point out the difficulty in eradicating supply, since wiping out production in one country leads to more production in other countries And they discuss a bit the arguments for legalization in various forms

Guns Violent crime and gun ownership both high in U.S. relative to Europe, correlated over time But causation unclear Maybe more guns cause more crime Or maybe more crime leads more people to want to own guns U.S., Canada, and Britain have similar burglary rates In Canada and Britain, about 50% of burglaries are “hot” (occur when victim is at home) In U.S., only 10% of burglaries are “hot” So gun ownership doesn’t seem to change overall number of burglaries, but does change the composition guns Cooter and Ulen also spend a little time discussing gun control They point out that both violent crime and gun ownership are high in the U.S. relative to Europe, and tend to be correlated over time as well However, causation could go either way more guns might cause more crime or more crime might lead more people to want to own guns Empirical work has not demonstrated a clear link between handgun ownership and violent crime One interesting tidbit The U.S., Canada, and Great Britain have similar burglary rates In the U.S., where handgun ownership is generally legal, a much smaller fraction of these burglaries occur when the victims are at home In the U.S., “hot” burglaries are only 10% of total burglaries In Canada and Great Britain, around 50% So widespread gun ownership doesn’t seem to reduce the overall number of burglaries, but it changes the composition

Racial profiling Well known that black drivers are more likely to be pulled over and searched than white drivers Jan 1995-Jan 1999, I-95 in Maryland: 18% of drivers were African-American, 63% of searches One explanation: police hate black people Different explanation: race could be correlated with other things that actually predict crime Maybe people in gold Lexuses with tinted windows and out-of-state plates are more likely to be carrying drugs Police stop more gold Lexuses with out-of-state plates More African-Americans drive gold Lexuses How to tell the difference? Need more information about what drivers were stopped/searched What to do if you don’t have data?

Knowles, Persico and Todd, “Racial Bias in Motor Vehicle Searches: Theory and Evidence” (2001, Journal of Political Economy) Game theory model of police and drivers Police get positive payoff from catching criminals, pay cost each time they stop a driver Drivers get some payoff from moving drugs, pay cost if they get caught Police may use lots of info (race plus other things) to determine who to pull over, but we (researcher) may not observe all that info

Knowles, Persico and Todd, “Racial Bias in Motor Vehicle Searches: Theory and Evidence” Suppose there’s a certain set of attributes (race plus other information) that make it very likely someone has drugs Then… Police will always search cars that fit those attributes So chance of being caught is very high for those drivers So drivers with those attributes will stop carrying drugs Only equilibrium is… For each set of attributes, some drivers carry, some don’t For each set of attributes, police search some cars, don’t search others

Knowles, Persico and Todd, “Racial Bias in Motor Vehicle Searches: Theory and Evidence” Mixed strategy equilibrium Police have to be indifferent between stopping and not stopping a given type of car/driver So if police see the cost as the same for all types… The payoff has to be the same, too Which means that in equilibrium, if police are not racist, the chance of finding drugs has to be the same for every car stopped Which means if we average over the other characteristics, black drivers should have the same probability of being guilty when they’re stopped as white drivers But this still might mean different search rates

Knowles, Persico and Todd, “Racial Bias in Motor Vehicle Searches: Theory and Evidence” Prediction: if police are not racist, each race could have different rates of being stopped/searched, but should have same rate of being guilty when searched Of drivers who were stopped in period they examine… 32% of white drivers were found to have drugs 34% of black drivers were found to have drugs Close enough to be consistent with “no racism” But only 11% of Hispanic drivers were found to have drugs, and only 22% of white women So police seemed to not be biased against black drivers, but to be biased against Hispanics, and white women? Other, more recent papers have found mixed evidence of racial bias

Crime in the U.S.

Crime in the U.S. As of 2005, over 2,000,000 prisoners, nearly 5,000,000 more on probation or parole Up from ~500,000 in 1980 93% male In federal prisons, 60% are drug-related Incarceration rate of 0.7% is 7 times that of Western Europe Cooter and Ulen estimate social cost of crime $100 billion spend annually on prevention and punishment 1/3 on police, 1/3 on prisons, 1/3 on courts, prosecutors, public defenders, probation officers, etc. Estimate another $100 billion on private crime prevention Estimate total social cost to be $500 billion, or 4% of GDP Cooter and Ulen try to estimate the social cost of crime The easy part to calculate is money spent on crime prevention and punishment, which is over $100 billion a year. One third is on police One third is on prisons One-third is on courts, prosecutors, public defenders, probation officers, and so on. They estimate nearly another $100 billion of private money is spent on crime prevention, such as alarms and security systems, private guards, and so on. They don’t really give numbers for the direct cost of crimes – stolen property, injuries, and so on They estimate the total social cost of crime to be $500 billion, or 4% of GDP

U.S. incarceration rate in context source: Center for Economic Policy Research http://www.cepr.net/index.php/publications/reports/the-high-budgetary-cost-of-incarceration/

U.S. incarceration rate in context source: Center for Economic Policy Research http://www.cepr.net/index.php/publications/reports/the-high-budgetary-cost-of-incarceration/

U.S. incarceration rate in context source: Center for Economic Policy Research http://www.cepr.net/index.php/publications/reports/the-high-budgetary-cost-of-incarceration/

U.S. spending on corrections source: Center for Economic Policy Research http://www.cepr.net/index.php/publications/reports/the-high-budgetary-cost-of-incarceration/

Crime in the U.S. Crime rates (relative to population) in the U.S… Decreased steadily from mid-1930s to early 1960s Increased sharply in 1960s and 1970s In 1980s, Nonviolent crimes committed by adults dropped sharply Violent crimes committed by adults dropped slightly Violent crimes by young people went up Violent and nonviolent crime rates dropped sharply in 1990s, continued to drop more slowly since 2000 Nonviolent crime rate in U.S. now similar to Europe Criminals in the U.S… Disproportionally young males Crime rates generally follow trend in fraction of population age 14-25 Both violent criminals and victims disproportionally African-American Relatively small number of people commit large fraction of violent crimes Tend to come from dysfunctional families, have relatives who are criminals… …do poorly in school, be drug- and alcohol-abusers… …live in poor/chaotic neighborhoods, and being misbehavior at young age source: Wikipedia http://en.wikipedia.org/wiki/Crime_in_the_United_States downloaded 11/28/2010

Crime in the U.S. source: Wikipedia http://en.wikipedia.org/wiki/Crime_in_the_United_States downloaded 11/28/2010

Why did U.S. crime rate fall in 1990s? What explains sharp drop in violent crime in U.S. in 1990s? Several explanations: deterrence and incapacitation decline of crack cocaine, which had driven much of crime in 1980s economic boom more precaution by victims change in policing strategies Donohue and Levitt give a different explanation: abortion (for change in policing strategies, see Gladwell “Tipping Point” among others)

Why did U.S. crime rate fall in 1990s? Donohue and Levitt U.S. Supreme Court legalized abortion in early 1973 Number of legal abortions ~ 1,000,000/year (compared to birth rate of 3,000,000) Violent crimes largely committed by males of certain ages Donohue and Levitt argue legalized abortion led to smaller “cohort” of people in high-crime age group starting in early 1990s The U.S. supreme court legalized abortion in early 1973 The number of legal abortions was on the scale of 1,000,000 a year (rising to 1.6 million by 1980), which is quite significant relative to a birth rate around 3,000,000 per year Violent crimes are largely committed by males of certain ages; Donohue and Levitt argue that legalized abortion led to a smaller “cohort” of people in the high-crime age group starting in the early 1990s. (A quick aside: birth rates, both total and relative to population, were indeed lower from 1973 to 1978 than they had been in previous years However, they had been dropping steadily well before 1973. Birth rates, both total and relative to population, were already substantially lower in the late 1960s and early 1970s than they had been in the 50s and early 60s Birth rate in 1972 was the lowest it had been since 1950, when population was much smaller. So even if the demographics of adolescent youth were a large part of the drop in crime, that’s not necessarily a direct result of legalized abortion.)

Why did U.S. crime rate fall in 1990s? Donohue and Levitt U.S. Supreme Court legalized abortion in early 1973 Number of legal abortions ~ 1,000,000/year (compared to birth rate of 3,000,000) Violent crimes largely committed by males of certain ages Donohue and Levitt argue legalized abortion led to smaller “cohort” of people in high-crime age group starting in early 1990s Evidence: Most of drop was reduction in crimes committed by young people Five states legalized abortion three years before Roe v Wade, saw drop in crime rates begin earlier States with higher abortion rates in late 1970s and early 1980s had more dramatic drops in crime from 1985 to 1997, no difference before Argue this explains 50% of drop in crime in 1990s Half of that from cohort size, half from composition Donohue and Levitt offer some evidence in support of their hypothesis First, most of the drop in crime rate was due to a drop in crimes committed by that age cohort – that is, the crime rates among “older” people did not fall in the 1990s Second, five states legalized abortion three years before Roe v. Wade; the drop in crime rates began earlier in those five states than in the rest of the country And states with higher abortion rates in the late 1970s and early 1980s did seem to have more dramatic drops in crime from 1985 to 1997, but no different crime patterns before that point Donohue and Levitt argue that legalized abortion explains 50% of the drop in crime in the 1990s of this 50%, half is due to its effect on the size of the cohort (just a reduction in the population of the high-crime age group) and the other half due to its composition: that those children not born were more likely to have been born to teenage mothers, single mothers, and the poor, and were therefore disproportionately likely to become criminals Obviously a controversial finding, and partially discredited – errors were found in the empirical work, some think the findings are wrong, others think the problems have been solved