Homo oeconomicus und (Wirtschafts-)Kriminalität? | Ringvorlesung "Das sogenannte Böse"

Summary

The lecture asks whether the economic model of homo economicus can help explain crime. Its answer is that rational-choice theory provides a useful framework, but it needs to be supplemented with evidence about bounded reasoning, emotions, social preferences, and real-world interaction.

1. Rational choice and crime

• In economics, homo economicus is a model of a decision-maker who chooses the best available option according to their preferences. The model does not require that a person be selfish; it focuses on how they respond to choices and incentives.
• Gary Becker applied this framework to crime. On this account, people may commit crimes when the expected benefit outweighs the expected costs, including the risk and consequences of punishment, and the value of alternative uses of their time and resources.
• Becker’s framework can also be used to assess crime policy. Society faces both the harms caused by crime and the costs of preventing and punishing it. Because eliminating crime entirely would require very costly measures, the model can imply an optimal crime rate above zero—not because crime is desirable, but because total social costs are being minimized.
• The model offers a systematic way to think about incentives, but its assumptions do not always match how people actually decide.

2. Behavioral factors that can shape criminal decisions

The lecture presents three cognitive factors and three emotional or social factors that can be added to a rational-choice analysis.

Cognitive factors

1. Risk perception and prospect theory

   • People do not necessarily treat probabilities as objective numbers. They may give small probabilities disproportionate weight while underweighting larger probabilities; extremely tiny probabilities may be treated as negligible.
   • This can affect how people respond to the perceived chance of being caught—for example, the risk of a tax audit.
   • The lecture also describes outcomes as being judged relative to a reference point. People tend to be more risk-averse over gains and more willing to take risks over losses. Since punishment is experienced as a loss, a person may sometimes prefer a risky outcome to a certain penalty.
   • Making the probability of enforcement clear and credible may therefore matter, not just increasing the severity of punishment.

2. Loss aversion and ownership

   • Losses can matter more to people than equivalent gains. This may make punishment more deterrent than a standard expected-utility model predicts.
   • An “endowment effect” experiment illustrates that people often value an item more once they own it: students who received university mugs were asked how much they would sell them for, while others were asked how much they would pay. Owners valued the mugs more highly.
   • Applied to theft, transferring an item from its owner to a thief may create more perceived harm than the item’s monetary value alone suggests. The implications for policy are ambiguous: greater perceived harm may justify more deterrence, while increased spending on private security might reduce the need for some public enforcement.

3. Time preferences

   • People often value immediate rewards more than equal rewards in the future and may discount future costs inconsistently.
   • This raises questions about how strongly long, delayed prison sentences deter crime. The lecture suggests considering whether shorter or alternative penalties, and reducing the delay between an offence and its punishment, might have greater influence on decisions.

Emotional and social factors

1. Anger and retaliation

   • The prospect of angry punishment can discourage wrongdoing and encourage cooperation.
   • But anger and retaliatory punishment can also provoke further retaliation and escalate conflict.

2. Shame, guilt, and social norms

   • Shame and guilt can motivate people to follow social norms even without an external reward.
   • The lecture cites tax-compliance experiments in which making tax evaders’ identities visible—for example, by publishing their photographs—encouraged compliance, apparently through shame.

3. Empathy, sympathy, and social bonds

   • Empathy can help people understand and share another person’s feelings; sympathy or social attachment can give another person’s welfare weight in one’s own decisions.
   • Positive relationships within a neighbourhood may support voluntary contributions to public safety and make harming neighbours less appealing, potentially reducing crime and the resources needed to control it.

3. Experimental studies of taking, trust, and theft

The lecture discusses experiments as a way to examine choices under controlled conditions.

• Dictator-game findings: In a basic version of the game, one participant can share money with another. The lecture reports that many participants give something, often invoking fairness. When participants are also allowed to take money rather than merely give or keep it, giving becomes less common and some take the maximum. In a version where the money first had to be earned, giving was lower still.
• Lesson about fairness: What people regard as fair may depend on the available actions. A person’s decision not to give may look different when taking from the other person is also an option.

Trust-game experiment with a theft option

• Two players begin with €12 each. The first player may send the second player €0–€6; the amount is multiplied by four. The second player can then return some of the resulting amount.
• In the theft version, the second player can also take money from the first player, up to an amount related to the first player’s investment.
• Under the standard self-interested version of the model, the second player would return nothing, so the first player would anticipate this and send nothing. The lecture notes that this outcome leaves potential gains from cooperation unrealized.
• In experiments, participants generally do invest and return money. In the lecturer’s study of 48 students, adding the theft option reduced the share of first players sending a positive amount: 67% in the theft version compared with 90% in the basic version. The average proportion returned relative to the amount invested was reported as broadly unchanged. In the theft version, all second players returned a positive amount; in the basic version, 82% did so.
• The lecture suggests reciprocity may help explain this result: the second player has already received something from the first player, which can encourage a positive response even when taking is possible.

4. Other research and overall conclusion

The lecture also mentions experimental work on reporting misconduct, lying, tax evasion, and tax manipulation. One reported pattern is that people were more likely to report misconduct when the victim was someone close to them; when the victim was in another group, reporting appeared more likely for a large theft than for a small one.

Overall, the rational-choice model helps organize questions about incentives and expected consequences, but it is not a complete description of behaviour. A fuller analysis of crime should also account for distorted risk perceptions, loss aversion, delayed consequences, emotions, social norms, and relationships. The lecture calls for further experiments involving interaction between people, rather than only isolated individual decisions.

Speakers and sources featured

• Main speaker: Prof. Dr. Claudia Keser, identified in the video metadata; the subtitles present the lecture without naming her.
• Audience member: An unnamed participant briefly responds during the discussion of the trust game.
• Researchers and sources discussed: Gary Becker; Daniel Kahneman and Amos Tversky; Franz van Winden and a co-author whose name is unclear in the auto-generated subtitles; and researchers involved in the dictator-game, trust-game, tax-compliance, and misconduct-reporting studies. The subtitles also refer to Göttingen colleagues and researchers including Melanie Dunger, Kubica, and Andreas Österreicher; several names and attributions are incomplete or garbled.