111

CHAPTER 16

DO POTENTIAL CRIMINALS CONTEMPLATE THEIR FATES?

Making Incentives Matter in the Face of Risk and Uncertainty

Early in the nineteenth century, England had more than 220 crimes, from picking pockets to murder, that would earn perpetrators one-way trips to the gallows. Thousands of criminals were executed before crowds of onlookers. Strangely enough, undeterred by the fates of their colleagues, pickpockets routinely worked the crowds at public hangings. The irony of this situation raises several questions of importance to modern policymaking: In what ways do punishments matter? To what extent can society address today’s crime problems with more severe punishments? And can we confidently assume that criminals are rational and informed? This chapter discusses crime, punishment, and approaches to uncertainty that are relevant to everyday decision making as well as to the choice to pick pockets at a pickpocket’s hanging.

They say crime never pays, but if that’s true, why do as many as 40 percent of automobile riders break mandatory seatbelt laws, 51 percent of drivers exceed the legal speed limit, and 79 percent of employees steal from their employers?¹ Classical economic theorists going back to Jeremy Bentham have argued that human behavior, including crime, is the result of rational calculations of the pleasure and pain that would result from various contemplated acts. In other words, even transgressions against society may be rooted in cost–benefit analysis.

¹ See www.investigation.com/articles/library/1999articles/articles22.htm.

112

Efficiency dictates that the optimal amount of lawbreaking is not 0. If you’re late for a career-launching interview on the Today Show, it might be wise to double park. If you can see lava from an erupting volcano in your rearview mirror, you might well exceed the speed limit. And if the only way to feed your family is to steal a loaf of bread, the benefits of that crime exceed the costs. Rational, informed criminals commit crimes until the marginal benefit from 1 more criminal act no longer exceeds the marginal cost. Unfortunately, criminals consider their own costs and benefits rather than those of society as a whole, so they commit more than the socially optimal quantity of crimes. Deadbolt locks, fences, and alarm systems that decrease the benefits of time spent committing crimes can sometimes forestall misdeeds, as can increased punishments or enforcement measures that raise the cost.

Are lawbreakers really that responsive to costs and benefits? Consider the example of speeding. If this infraction is rational, we should expect the frequency to increase among those who receive greater benefits from it. If there is no accident or ticket, the benefit from speeding is a saving of time. The opportunity cost of time spent driving is higher for those who earn more per hour, so economic theory would predict more speeding among those with higher incomes. Indeed, the 2005 Northwest Insurance Poll by PEMCO Insurance found that two-thirds of drivers with annual incomes exceeding $75,000 speed often or sometimes, compared to 55 percent of drivers who earn between $35,000 and $74,999, and only 43 percent of those who make less than $35,000.

On the cost side, other things being equal, we would expect larger fines to result in fewer infractions. The relationship between fines and infractions is more complex than that, however, because the value of fines may not be known and because the rational driver must also consider the probability of being caught and the large effect that speeding tickets may have on insurance costs. Concerned citizens argue that differing risks of being caught make it rational to drive more aggressively in Oregon, which has 241 state troopers, than in Washington state, which has 658 troopers, or in Florida, which has 1,360.² Even if people respond rationally to what they know well, such as the opportunity cost of our time, issues of uncertainty become important to decision making and policy. Most people do not know the number of troopers in their states or the fines for driving 20 mph over the speed limit. The influence of uncertain aspects of law enforcement depends on whether a person is blissfully ignorant of strict laws or paralyzed by the fear of potentially high punishments. The remainder of this chapter explains some rational ways to approach the uncertainties in our lives, the extent of exceptions to rationality, and the implications for public policy.

² See www.blueoregon.com/2005/04/oregon_the_stat.html.

Criminals are not the only ones grappling with uncertainty. When developer Donald Trump contemplated building the Trump World Tower in New York City, the world’s largest residential tower, he did so not knowing the future of housing prices. Students face uncertainty about their employment prospects after graduation. Consumers are uncertain about the quality and safety of many of the products available for purchase. Business owners face uncertainty about their future levels of sales. The insurance industry exists to reduce uncertainty about the costs of health care, car and home ownership, and a host of misfortunes. As Benjamin Franklin put it, “In this world nothing is certain but death and taxes,” and there are even insurance markets to deal with the uncertainties of death (euphemistically called life insurance) and tax audits.

113

Expected-value calculations offer a rational way to deal with risk. The expected value of an outcome is calculated by multiplying the probability of that outcome by its value. Consider the outlook for a rational and informed potential auto thief in Minnesota who would weigh the benefit he or she receives from a stolen car against the expected punishment. About 21 percent of car thieves in Minnesota are arrested, 91 percent of offenders are incarcerated, and the average prison sentence is 27 months.³ The average income in Minnesota is close to $3,000 per month, so the 27 months of incarceration would amount to an income loss of 27 × $3,000 = $81,000. If this were the only loss associated with incarceration, the expected value of punishment for stealing a car would be 0.21 × 0.91 × $81,000 = $15,479, and a car worth $20,000 would be enticing. That helps to explain why more than 15,000 cars are stolen in Minnesota in a typical year. Many people would lose far more than $81,000, considering the difficulty of finding a job after serving time, separation from family, a tarnished reputation, and extreme embarrassment.⁴ Conversely, other people have little to lose in terms of employment, family, or reputation. As a result, some people are avid car thieves and most people wouldn’t even consider committing the crime.

³ See www.msgc.state.mn.us/Data%20Reports/mv1998.pdf.

⁴ Economist John R. Lott has studied the importance of these aspects. See “The Optimal Level of Criminal Fines in the Presence of Reputation,” Managerial and Decision Economics (July–August 1996), 363–380; and “An Attempt at Measuring the Total Monetary Penalty from Drug Convictions: The Importance of an Individual’s Reputation,” Journal of Legal Studies (January 1992), 159–187.

Risky decisions are influenced by your attitude toward risk. Even if the theft of a car would net you more than the expected value of punishment, the downside risk of a heavy fine or imprisonment might be too great for you to accept. A person who cares only about the expected value of an outcome is described as risk neutral. Risk-averse people care about variation in possible outcomes in addition to the expected value and would pay extra to avoid uncertainty. For example, a risk-averse person would be willing to pay more than the expected value of a tax evasion fine for insurance that would pay any fines resulting from a tax audit. Risk-loving people prefer variation in possible outcomes. These are the folks who go to Las Vegas and pay more than the expected outcome (as is generally the case in gambling situations) for the possibility of winning big. Economists often simplify analysis by assuming that decision makers are risk neutral, which is likely for larger firms and wealthy individuals who can handle most downside risks and close to the truth for the average individual, although most people lean toward risk aversion.

When contemplating infractions of the law, potential criminals face uncertainty about being caught, being convicted if caught, and receiving various sentences if convicted. The U.S. Department of Justice reports that the probability of being arrested is 67 percent for murder and nonnegligent manslaughter, 58 percent for aggravated assault, 52 percent for forcible rape, 47 percent for all violent crime, and 18 percent for all property crime. Felony conviction rates are 73 percent for drug trafficking; 67 percent for murder, burglary, and weapons offenses; and, at the low end, 41 percent for assault. Seventy-one percent of all convicted felons are sentenced to periods of confinement, and 41 percent are ordered to pay fines or victim restitution, receive treatment, perform community service, or comply with some other additional penalty.

114

The next section discusses punishment and the likelihood that various types of criminals are cogent enough to be deterred by changes in the penal code.

There are four primary reasons for punishing criminals:

According to the deterrence hypothesis, an increase in the punishment for a crime—effectively, a rise in the price of committing the crime—will decrease the quantity of crimes committed. The validity of the deterrence hypothesis depends on people being sufficiently informed and rational to make the necessary calculations. In order to respond appropriately to harsher punishments, such as the death penalty or mandatory sentencing, people must have some idea of the punishments for contemplated crimes and of the likelihood of being caught, and they must consider those possibilities when deciding whether to act.

A research team and I interviewed 278 male inmates at two medium-security state prisons and one county jail to determine how many of them possessed the information and the mindset needed to respond to punishments at the times of their offenses.⁵ Sixty-three percent of the criminals either did not think they would be apprehended or did not think about the possibility. Fifty-three percent either did not know what the punishment would be if they were convicted or did not think about it. And a combined 76 percent of the inmates did not contemplate their fates, had no meaningful information with which to do so, or both. Still others had inhibited responses to known punishments as a result of rage, drugs, psychosis, or heat-of-the-moment impulses.

⁵ See David A. Anderson, “The Deterrence Hypothesis and Picking Pockets at the Pick-Pocket’s Hanging,” American Law and Economics Review (Fall 2002), 4:2, 295–313.

Current threats of punishment deter countless would-be criminals; the relevant question here is whether harsher punishments would deter many of those who are active criminals despite existing disincentives. The ability to respond to punishments differed among perpetrators of various types of crimes. Whereas 89 percent of those incarcerated for escape, parole violations, nonpayment of child support, and forgery had thought about the punishments they might receive, only 45 percent of murderers had contemplated likely punishments. Between 21 and 47 percent of nonviolent offenders knew exactly what their punishments would be, but only 4 to 16 percent of the violent criminals could predict their sentences. As for the possibility of thwarting existing criminals with stricter laws and increased detection rates, almost half the people who committed minor offenses were calculating and informed enough to be deterred by a change in law enforcement. Unfortunately, 89 percent of deadly criminals, 91 percent of sex offenders, and 88 percent of robbers lacked the requirements for making informed, rational judgments and for responding to harsher punishments. Given these criminals’ mindsets and the information available to them, simply ratcheting up punishments is unlikely to have the intended effect, and supplemental approaches are advisable to better address the worst types of crimes.

115

When criminals are lucid and well informed, as is more often the case in property crimes than in violent crimes, deterrence can result from an increase in the expected cost of crime. For efficiency, we want the expected cost of each crime to equal its cost to society. That way people will commit the crime only if their benefit exceeds the burden to society. For instance, if the illegal dumping of old tires causes $500 worth of ugliness and cleanup costs for a community, a cost (including the expected value of the fine and any costs associated with guilt or risk taking) of $500 will prevent tire dumping except in those few cases in which the benefit is greater than the $500 cost to society. When the benefit truly exceeds the cost, from an efficiency standpoint, it makes sense for the dumping to occur.

Because the expected value of a fine is the product of the likelihood of arrest and conviction multiplied by the actual fine, it’s noteworthy that any desired increase in the expected fine can be achieved by increasing any of these three elements. In light of the large expenditures on policing and trials needed to increase arrest and conviction rates, the cost-minimizing law enforcement strategy is to maximize the punishment and thereby minimize the arrest and conviction rates necessary to achieve a particular expected cost. Suppose an expected fine of $500 yields the efficient amount of tire dumping. This could be achieved with a 50 percent probability of arrest and conviction and a $1,000 fine (0.50 × $1,000 = $500) or a 1 percent probability of paying a $50,000 fine (0.01 × $50,000 = $500), but the policing cost of catching 1 percent of perpetrators would be far lower than that of catching 50 percent.

Of course, the cost-minimizing enforcement plan comes down hard on the few who do get punished. Chapter 13 introduced risk aversion and related risk and guilt burdens. The possibility of having to pay $50,000 creates a burden for risk-averse people, who prefer as little variation in the potential outcomes as possible. If the risk burden (the amount that risk-averse people would pay to avoid uncertainty about the possible outcome) is $100 and there is no guilt burden, an expected fine of $400 will yield efficient behavior because the total cost of dumping, $100 + $400 = $500, equals the social cost. If the expected fine is $500, then the total cost of $100 + $500 = $600 will deter people from dumping even when their benefit is, say, $550, which exceeds the social cost and makes the dumping efficient. It may seem strange to hear that it could be efficient to dump tires, but remember that the precondition is that the combination of every related cost falls below the benefit received.

For those potential criminals who are less informed or perhaps less rational, including the most violent offenders, effective deterrence requires policies beyond adjustments in the expected punishment cost. One approach is to increase the level of information and prudence among potential criminals. Educational programs could foster better methods of decision making. For those who feel invincible, billboards touting “There are a million people in prison who didn’t think they’d be caught” might be a start. For those who don’t know the punishments for crimes they consider, a campaign advertising specific consequences would help. And for those who might snap emotionally, teachings on how to handle strong emotions could make a difference in a way that ratcheting up punishments cannot.

116

There is also a more general deterrent founded on economic theory and proven in the real world: education. Remember the importance of opportunity costs, which include the next-best alternative value of time spent on crime or in prison. Those with more education have more to lose by committing crimes because they are more likely to be gainfully employed and to earn higher incomes than those with less education. The frequency of most types of crime dropped steadily in the 1990s as employment prospects improved and rose in the early 2000s when more people were out of work and had a lower opportunity cost of time spent on crime and punishment.

Although some crimes create a net gain for society, most are inefficient, either because of the criminals’ neglect of the costs imposed on others or because of poor information, powerful emotions, or perceived invincibility on the part of the criminals. A perfunctory application of economic principles to fight crime may fail to place effective incentives in the right places. The deterrence hypothesis that harsher penalties will deter crimes is valid only when potential criminals have some idea of the penalties for contemplated crimes and think they might be caught. Most criminals know neither the penalty for laws they break nor the risk of being caught, so increases in fines or enforcement lack the desired punch. Alarmingly, the information and mindset required to respond to typical crime prevention efforts, such as mandatory sentencing and “three-strikes” laws with strict penalties for third offenses, is particularly lacking among the most violent offenders. Effective policies against the worst types of crime must address the information costs that prevent criminals from making appropriate decisions about risks of detection, conviction, and punishment. Opportunity costs are relatively well known, which points to remedies such as job creation, drug rehabilitation, and education—services that elevate careers and reputations among more certain and influential opportunity costs of a life of crime.

⁶ For example, see David Anderson, Sometimes I Get So Angry! Anger Management for Everyone (Danville, KY: Pensive Press, 2007).