⚠️Automatic translation pending review by an economist.
For the uninitiated, studying crime using economic tools may seem confusing: banditry is often associated with irrational behavior, whereas economics is dedicated to the study of rational decisions.
However, since (at least) Plato(The Republic, Book 2) and the fable of Gyges’ ring, we know that the decision to engage in illegal (or unjust, in the fable) activity depends on the probability of being arrested, which is itself a function of an individual’s skills (his ability to bend the law, in Plato’s text[1]); engaging in criminal activity is therefore the result of a cost-benefit calculation in which an individual judges the potential gain (monetary, psychological, etc.) of a crime against his or her risk.) of a crime against its potential cost (a potential conviction, but also the opportunity cost of the crime). Similarly, if we follow Hans Magnus Enzensberger in his book Crime and Politics (1964), the decisions taken by Al Capone and his crime syndicate can be analyzed as those of a company seeking to acquire monopoly power in its market[2].
Examples and analogies could be multiplied; the point to remember is that criminals act according to a number of perfectly rational forces and incentives. Economic science can shed light on this phenomenon, and has done so since Gary Becker (1968), who defined the gains and costs of crime and deduced the optimal level of security a society should choose.
The approach is not as absurd as it may seem. But what can economics really teach us about crime? They are vast, but in a nutshell, economics is concerned with choices and interactions[3], while the other sciences are interested in profiles and environments. Criminology, psychology or sociology can determine the factors that encourage people to embark on a career in crime (in economic terms, the factors that will determine an individual’s « preferences » for crime), whereas economics takes these factors as given and is interested in choices. The difference may seem minimal, but the consequences are not: when we are able to understand the choices made by individuals, it is then possible to estimate how a population will react if the parameters of this choice are modified (i.e. to carry out a counter-factual analysis). For example, it becomes possible to quantify the number of additional crimes committed if poverty or unemployment[4] increase, if prison sentences or the number of police officers in a region are modified, etc. One of the primary interests of the economic approach to crime is therefore to enable a clear analysis of public policies aimed at combating crime.
The remainder of this article is therefore devoted to an analysis of a number of policies or received ideas for reducing crime, using economic reasoning and econometric analysis. The aim here is not to be exhaustive, but rather to illustrate how economists can help combat crime.
Increasing police forces
The first policy that comes to mind is simply to increase police resources and manpower in a particular area (priority security zones, for example); while this may seem natural, there is no guarantee that it will be effective. For this to be the case, police reinforcements must lead to a reduction in crime, and this reduction must be sufficiently large for the measure to be socially effective (i.e. its benefit in terms of crime reduction must outweigh its cost).
The fact that police reinforcements reduce crime seems intuitive, and is also readily accepted in economic theory[5]. Empirically, however, the issue is much more complicated: as the number of police officers in a city is set primarily as a function of crime in that city, the correlation between the two variables is likely to be positive (in technical terms, there’s an endogeneity problem). However, we can take comfort in the fact that the most recent studies show that police reinforcements do indeed reduce crime. Stephen Machin and Olivier Marie (2005), for example, have shown that the Street Crime Initiative in England prevented over 10,000 thefts, and that one pound spent on this program yielded between four and a half and five pounds (in terms of monetary benefit per theft prevented).
The study also confirms a certain « rigidity » in crime: increasing police forces in one area does not seem to lead criminals to move to a neighboring area. This is an important finding, as it lends a certain legitimacy to targeted actions. They have a real impact on crime, and don’t just displace it[6].
Weapons control
The ease with which firearms can be obtained is often seen as having a direct impact on the ease of committing crimes, particularly homicides. So it seems tempting to argue that greater gun control would lead to fewer murders. In fact, this is the argument often heard in France in the aftermath of a massacre in the United States. The reality is far more complex, however, since firearms serve both as an offensive technology (enabling crimes to be committed) and as a defensive technology: a weapon can be used when attacked, and some criminals may hesitate to commit an assault or burglary (etc.) knowing that their victim may be armed.
This means separating the population into two categories: criminals (who can use a weapon either offensively or defensively) and non-criminals (who can only use a weapon defensively). At least in the short term, controlling gun sales in areas where there is already a large stockpile (e.g. the USA) is likely to make gun ownership more difficult for non-criminals than for criminals (the latter often obtain their guns through burglary or by buying stolen weapons, so they will be indirectly affected by the measure). This control policy will therefore skew the gun ownership ratio in favour of criminals and encourage crime; this is why theoretical and empirical studies on the subject do not put forward gun control as a way of reducing crime or homicides (Kovandzic, Schaffer Kleck (2008), Ehrlich Saito (2010)).
If the available stock of firearms is initially zero (or close to it), allowing the free sale of weapons has a detrimental effect. It’s a tautology, but when nobody is armed, criminals aren’t either, so the ratio of gun ownership between criminals and non-criminals is at its « second best ». The « best situation » (or « first best ») would be to distribute guns to non-criminals when criminals are unarmed, in order to deter them from committing a crime. Of course, this « best situation » is not possible, because we can’t distinguish a criminal from a non-criminal, and we’d have to make sure that the weapons possessed by the latter would not eventually be recovered by the former. Second best » is therefore the best situation that can be achieved.
Play on prison sentences
One fact that the theory of crime as rational choice still fails to explain clearly is why the severity of sentences seems to have no influence on crime. If criminals weighed the gain of a crime against its « expected cost » (roughly speaking: the probability of being convicted multiplied by the severity of the sentence), then the harsher the punishment, the lower the incentive to commit criminal acts. Yet several empirical studies show that the probability of being caught has a large influence on crime, while the severity of sentences has almost none[7]. On the contrary, criminals who have already been convicted have a much higher probability of committing a crime again (i.e., the problem of recidivism) than non-criminals; they also tend to commit new crimes that are more serious than old ones. This finding is largely influenced by prison conditions. The harsher the prison conditions, the more likely convicts are to re-offend and commit new, more serious crimes (M. Keith Chen and Jesse M. Shapiro (2002)).
Numerous theories attempt to explain this fact, appealing for example to the transmission of skills in prison, to the taste for violence that criminals can develop (sometimes unwittingly); due to a loss of reference points or opportunities on the job market following imprisonment (crime then pays all the more in relation to legal work), etc. These questions deserve to be shed light on by a wide range of experts. These questions deserve to be examined in much greater depth, but we won’t go into them here.
However, since increasing prison sentences does not reduce crime, we may be tempted to advocate the opposite policy: reduce imprisonment as much as possible, so that criminals can reintegrate more easily, acquire fewer « skills », and also reduce the (not insignificant) cost to society of imprisonment. What do we think of this proposal?
First of all, we need to understand the value of prison sentences as understood by economists. The primary aim of sentencing is not to punish the criminal (indeed, if sanctions are optimally determined and a crime is committed, it’s because the crime was « profitable » for society[8]), but to constitute a credible threat so that crimes that are not profitable for society are not committed. In practice, however, it is well known that socially harmful crimes will be committed, so a prison sentence may also have the (secondary) aim of preventing an individual from doing harm for the period during which he or she is behind bars (see, for example, Polinsky and Shavel (2007) for an introduction to » Optimal Deterrence » theory, or Miceli (2009) for an « incapacitation » model). A good policy must therefore take both objectives into account.
With this in mind[9], Yeh (2010) proposes an interesting policy: that each sentence (handed down in the USA) should be divided into two equal periods of time; the first spent in prison and the second spent at home, with an electronic bracelet[10]. The general idea is as follows: the first part of the sentence is a real prison sentence, so it remains a sufficient deterrent. The second part of the sentence not only allows for a better readjustment to civilian life, but also drastically reduces the cost to society of having a man in prison.
Now for the second aspect: incapacitation. Yeh draws on a study by Padgett, Bales and Blomberg (2006), which found that 94.7% of electronic bracelet wearers respected the terms of their sentence (i.e., they neither committed a new crime nor violated the conditions of their probation). The conclusion of Yeh’s cost-benefit analysis is edifying: for every dollar spent, the State would save over twelve, for a social value of crimes avoided of over 480 billion. Finally, Yeh points out that his study does not internalize many of the costs associated with sentencing (e.g. loss of employability), so the real benefit could be even higher.
Conclusion
Of course, the three policies analyzed above cannot be replicated as they stand. Even so, replicating these public policies would require much more discussion and analysis. They do, however, introduce how economists can use decision theory, econometrics and cost-benefit analysis to help fight crime more effectively.
Notes
[1]The fable explores a borderline case: that of an invisible criminal; the probability of being arrested or tried is (implicitly) set at zero for anyone holding Gyges’ ring.
[2]Hans Magnus Enzensberger is not an economist, however, and his analysis lacks precision. His idea was later taken up, however, and organized crime is now studied as an industry.
[3]The study of the interactions between criminals, victims and police is widely studied through game theory and will be little discussed here; the interested reader can, for example, read the literature review by Bill McCarthy (2002).
[4]However, it seems that the relationship between crime and unemployment is positive overall (e.g. Lee and Holoviak (2006)), but not as obvious as it might seem at first glance; take the example of burglary: an increase in unemployment leads to fewer opportunities on the labor market, and therefore increases the incentive to engage in criminal activity. However, according to Allen (1996), an increase in unemployment increases the number of people staying at home during the day, making opportunistic burglary more difficult.
[5]However, economic analysis also warns against over-simplistic relationships. More reinforcements can lead to a reduction in the « effort » made by police officers already on the scene. Here, the term « effort » can be interpreted in different ways: if the objective is to maintain a satisfactory level of crime, then police reinforcements enable the objective to be achieved with less effort per police officer. However, a « less negative » interpretation is possible: if police reinforcements enable a reduction in the overtime worked by each officer, then with more staff the same result on crime will be achieved (only the pace of work has changed).
[6]Nor does targeted action on a certain type of crime (e.g. theft, delinquency, etc.) seem to cause criminals to change their activity (from theft to drug trafficking, for example). But targeted actions have no externalities either: a campaign against theft will have little or no effect on drug dealers, for example.
[7]In fact, rational choice theory is not totally disarmed either. If an individual gives more weight to the present than to the future, and is interested in the « net present value » of a crime, each additional year in prison will have less impact than the previous one. Thus, increasing sentences can only have a very marginal impact on the number of crimes committed.
[8]A good example of this is speeding fines. Driving too fast increases the number of deaths on the roads, and therefore causes « damage » to society. But it can happen that « the man in the street » has an urgent appointment that requires him to violate them. In the optimum case, he must break the speed limits if, and only if, the benefit he derives from arriving on time for his appointment is greater than the damage he will inflict on society by driving too fast. The fine must make him internalize (take into account in his calculations) the harm he is doing, and not necessarily make him renounce his action.
[9]This policy also aims to relieve prison overcrowding.
[10]In reality, almost no-one spends their entire sentence in prison. However, there is a high rate of recidivism among those who are released before the end of their sentence; electronic monitoring, if generalized as proposed by Yeh, would not only avoid the cost of detention but also prevent crime.
References
T. Kovandzic, M. Schaffer and G.Kleck, Estimating the causal effect of gun prevalence on homicide rates: a local average treatment effect approach, IZA discussion paper 2008.
Hans Magnus Enzensberger, Politique et Crime, 1964. French translation by Edition Gallimard, 1967.
I.Ehrlich T.Saito, Taxing Guns VS. Taxing Crimes: an application of the market for offenses model, NBER working paper 2010.
S. Machin, O. Marie, Crime and Police Resources: the Street Crime Initiative, CEP Discussion Paper 2005.
G. Becker, Crime and Punishment: An Economic Approach, Journal of Political Economy, 1968.
T. Miceli, Deterrence and Incapacitation: Towards a Unified Theory of Criminal Punishment, Working Paper 2009.
M.Polinsky, S. Shavell, The Theory of Public Enforcement of Law, Handbook of Law and Economics, vol 1, 2007.
K.g. Padgett, W.D. Bales, T.G. Blomberg, Under Surveillance: an empirical test of the effectiveness and consequences of electronic monitoring, Criminology and Public Policy, Vol5, Issue1, 2006.
S.S. Yeh, Cost-benefit analysis of reducing crime through electronic monitoring of parolee and probationers, Journal of Criminal Justice, vol 38, 2010.
D. Lee and S.J. Holoviak, Unemployment and crime: an empirical investigation, Applied Economics Letter, V13, 2006.
B. McCarthy, New Economics of Sociological Criminology, Annual Reviews of Sociology, 2002.
M. K. Chen, J. M. Shapiro, Does Prison Harden Inmates? A discontinuity-based Approach, Working paper, 2002.
Hans Magnus Enzensberger, Politique et Crime, 1964. French translation by Edition Gallimard, 1967.
I.Ehrlich T.Saito, Taxing Guns VS. Taxing Crimes: an application of the market for offenses model, NBER working paper 2010.
S. Machin, O. Marie, Crime and Police Resources: the Street Crime Initiative, CEP Discussion Paper 2005.
G. Becker, Crime and Punishment: An Economic Approach, Journal of Political Economy, 1968.
T. Miceli, Deterrence and Incapacitation: Towards a Unified Theory of Criminal Punishment, Working Paper 2009.
M.Polinsky, S. Shavell, The Theory of Public Enforcement of Law, Handbook of Law and Economics, vol 1, 2007.
K.g. Padgett, W.D. Bales, T.G. Blomberg, Under Surveillance: an empirical test of the effectiveness and consequences of electronic monitoring, Criminology and Public Policy, Vol5, Issue1, 2006.
S.S. Yeh, Cost-benefit analysis of reducing crime through electronic monitoring of parolee and probationers, Journal of Criminal Justice, vol 38, 2010.
D. Lee and S.J. Holoviak, Unemployment and crime: an empirical investigation, Applied Economics Letter, V13, 2006.
B. McCarthy, New Economics of Sociological Criminology, Annual Reviews of Sociology, 2002.
M. K. Chen, J. M. Shapiro, Does Prison Harden Inmates? A discontinuity-based Approach, Working paper, 2002.
