Article Review

Crime policy and behavioural economics: an article review

The variety of posts of this blog is the proof that the potential applications of Behavioural Economics are plenty. The present article reviews a paper in which the author, Pickett (2018), illustrates some examples of its use for crime deterrence. The decision of committing a crime has been analysed in economic terms with a simple model where potential offenders evaluate the expected benefits and costs of breaking the law. The expected benefits depend on the nature of the crime, while the expected costs are given by the perceived probability of being arrested and the cost attached to being sentenced to prison. Classical ideas to improve deterrence are related to increasing expected costs; however, this is often too costly (increasing surveillance, police checkpoints, etc.). Behavioural economics can intervene to find effective policies that are much less expensive. This perspective is not widely adopted and, thus, the article is intended to stimulate future research in this context. In particular, the author discusses three experiments to support his position.

In the first experiment, the participants were shown three different videos where a policeman was talking about drunk driving and law enforcement. The first one was used as control, while the other two were either an “unpacking video” (listing all the possible ways in which one could be arrested if found driving under the influence of alcohol) or a “pseudo-certainty video” (declaring explicitly that the probability of being arrested for drunk driving conditional on encountering the police is nearly 100%). Then, subjects were asked to evaluate the perceived probability of being arrested and, in particular, the fear of being arrested (which seems to be a much more important factor contributing to deterrence) if driving after drinking alcohol. The results showed little difference between the control group and those exposed to the “unpacking video”. However, who saw the “pseudo-certainty video” were assigning a higher probability to the arrest and were much more fearful. In the article, this is explained through the anchoring heuristic: priming the subjects with a video reporting a conditional probability of 100% increases the perceived probability of the event itself (not conditioned on having encountered the police). Even though, in general, the perceived risk of being arrested is higher than the actual one, from a policy perspective, advertisement informing about the high probability of being arrested conditional on encountering a police checkpoint would further increase the expected costs (and, therefore, lower the probability) of committing a crime.

In the second experiment, the author wanted to understand whether situational factors may influence crime rates. In order to test this, he showed pictures of women with a purse, either big or small. Subjects were then asked to guess the total value (in dollars) contained in the purse and to rank how easy it would have been to steal it. The results were not conclusive, as the bigger purse was judged to contain more value and, at the same time, harder to steal. Thus, it is difficult to understand which one would be the preferred target of a criminal. However, it is clear that a situational factor (such as the size of the purse) influences the judgement and, therefore, the decision of the criminal: the author encourages to study this, extending the analysis to other factors that might be relevant.

The last experiment was related to the common belief that deterrence depends on the length of the punishment. The participants were asked how afraid they were if given a certain punishment, which ranged from 6 months to 8 years. Even though the responses were different for males and females, none of the two groups expressed fear increasing with sentence duration: quite surprisingly, sentences of six months and eight years induce similar levels of fear. This suggests that the Prospect Theory, where concepts such as loss aversion and reference dependence are accounted for, is useful for analysing the behaviour of criminals: for sentences above six months, people are somehow indifferent, meaning that the disutility they get from a sentence of eight years and one of six months is at least comparable (if referring to the classical utility function described by the Prospect Theory, it should be more or less flat for sentences longer than six months). This has very interesting policy implications, especially since, most of the time, policy makers believe that more severe punishment will reduce the incentives to offend.

These three experiments have clear limitations. One above all, the subjects were university students which hardly (hopefully!) had experience as criminals. For instance, one could argue that the judgement about the duration of the sentence may be affected by past experiences in prison. Or, perhaps, we could say that these experiments regard minor crimes and it is also legitimate to be suspicious about the possibility of using behavioural economics to deter criminals from committing major crimes (a homicide might derive from very peculiar reasons and thus it is hard to think that nudges could play a role in preventing it). As a matter of fact, the author is aware of the inconclusiveness of his results. Nevertheless, the point that should be particularly underlined is that research in this field is still emerging and there is room for improvement. It is reasonable to think that studying crime through the lenses of behavioural economics might produce fruitful results.


Pickett, J. T. (2018). Using behavioral economics to advance deterrence research and improve crime policy: Some illustrative experiments. Crime & Delinquency, 64(12), 1636-1659.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s