首页 | 农村土地改革 | 法经济学茶座 | 法经济学论坛
     网站首页≯ 刑法≯
刑事行为的经济学(1)

 ECONOMICS OF CRIMINAL BEHAVIOR

Erling Eide

Professor, Faculty of Law, University of Oslo

© Copyright 1999 Erling Eide

 

Abstract

Models of criminal behavior, where a person is assumed to act rationally on the

basis of costs and benefits of legal and illegal opportunities, are presented in

this chapter. Most of these models are similar to models of portfolio choice and

of supply of labor. The empirical studies that are surveyed use various types of

regression analyses and employ data from states and police regions down to

campuses and individuals. Most studies corroborate the hypothesis that the

probability of punishment, and to a lesser degree also the severity of

punishment, has a deterrent effect on crime. The effects of various economic

factors are less clear, although unemployment seems to increase crime.

Methodological problems relating to the assumption of rationality, to statistical

identification of equations, to measurement errors, and to operationalization of

theoretical variables are discussed.

JEL classification: K42

Keywords: Crime, Punishment, Sanctions, Rational Choice

 

1. Introduction

 

Theories of crime are abundant. Various mental, physical, developmental,

economic, social, cultural, and other causes have been launched as explanations

of why people offend. Concepts like depravity, insanity, abnormality, deviance

and deprivation are used to characterize those who commit crimes. During the

last 30 years economists have invaded the field using their all-embracing model

of individual rational behavior, where a criminal act is preferred and chosen if

the total pay-off, including that of sanctions and other costs, is higher than that

of legal alternatives. Offenders are primarily not regarded as deviant

individuals with atypical motivations, but rather as simple, normal persons like

the rest of us. The theory of deterrence thus obtained is regarded as nothing but

a special case of the general theory of rational behavior under uncertainty.

Assuming that individual preferences are constant, the model can be used to

predict how changes in the probability and severity of sanctions and in various

socioeconomic factors may affect the amount of crime. Even if most of those

who violate certain laws differ systematically from those who abide by the same

laws, the former, like the latter, do respond to incentives, that is, to sanctions

and economic conditions. Empirical tests with increasing statistical rigour and

refinement have been carried out on the basis of this theory.

 

Whereas the general preventive effects of sanctions for a long time have

occupied a main position in penal legislation and sentencing policy, such

effects were almost totally neglected in criminology and modern sociology until

the late 1960s. Criminologists have been more interested in rehabilitation and

treatment, and many are still reluctant to accept studies of deterrence in general

and models of criminal behavior based on rational choice in particular.

However, scholars who are reluctant to accept the assumption of rational

choice, still find interest in the rather rigorous empirical studies in the

economics of crime literature (Andenaes, 1975), and sociologists have in recent

years been inspired to carry out similar research. Below, mainly studies made

by economists are included.

 

2. The Basic Model

 

Theories of criminal behavior based more or less on the assumption of rational

choice were proposed by Beccaria and Bentham. Bentham ([1788] 1843, p.

399) wrote that ‘the profit of the crime is the force which urges man to

delinquency: the pain of the punishment is the force employed to restrain him

from it. If the first of these forces be the greater, the crime will be committed;

if the second, the crime will not be committed.’ From the beginning of this

century interest in their point of view dwindled as a plethora of other theories

were developed. The main idea of Bentham was vitalized and modernized in

the pathbreaking article on Crime and Punishment by Becker (1968), who

suggests that ‘a useful theory of criminal behavior can dispense with special

theories of anomie, psychological inadequacies, or inheritance of special traits

and simply extend the economist’s usual analysis of choice’ (p. 170). He argues

that criminals are like anyone else, and assumes that an individual behaves as

if he is a rational utility maximizer. As the total outcome of a criminal act is

uncertain, Becker employs the usual assumption that people act as if they were

maximizing expected utility, and also that utility is a positive function of

income. The individual’s expected utility E[U] from committing an offense is:

 

 

where is the individual’s von Neumann-Morgenstern utility function, P is

the subjective probability of being caught and convicted, Y is the monetary plus

psychic income (that is, the monetary equivalent) from an offense, and f is the

monetary equivalent of the punishment. The individual will commit the offense

if the expected utility is positive, and he will not if it is negative. The common

assumption of stable preferences provides a solid foundation for generating

predictions about responses to various changes in parameters and, according

to Becker, prevents the analyst from succumbing to the temptation of simply

postulating the required shift in preferences in order to ‘explain’ all apparent

contradictions to his predictions. Analysis of comparative statics shows that

increases in either the probability or the severity of punishment might change

the expected utility from being positive to being negative. For society as a

whole Becker introduces a ‘supply of offense function’, where the two factors

have an effect on the total amount of crime.

 

Whereas Becker considers the income and punishment equivalents of an

offense separated from other income, later authors, in accordance with Brown

and Reynolds (1973), take the individual’s initial income position as a point of

reference. Expected utility becomes

 

 

where W is present income and g is gains from crime. Here, the crime will be

committed if the expected utility is higher than the utility of the initial income

W. Furthermore, it is sometimes assumed that the offender in case of conviction

might retain some gain from the offense. Becker demonstrated that if the

elasticity of the expected utility with respect to the probability of punishment

exceeded the elasticity of the expected utility with respect to conviction (both

in absolute values), the offenders were risk lovers. Empirical studies by Becker

and others corroborated this result. As shown by Brown and Reynolds (1973)

equation (2), at variance with equation (1), does not imply such a conclusion.

 

3. Extensions of the Basic Model

 

Later, several types of economic models of crime have been developed, all of

which draw on the theory of supply and the theory of behavior towards risk.

The simplest one is very similar to models of portfolio choice, where a person’s

wealth is allocated between various risky and non-risky projects. In the

economics of crime version of this model the illegal alternatives are considered

as risky mainly because of uncertainty about punishment. Allingham and

Sandmo (1972), Kolm (1973), and Singh (1973) have constructed such models

for tax evasion, where the individual is confronted with the problem of deciding

what proportion of income not to report to the tax authorities. At variance with

Becker’s model where the income of crime is a parameter, here the income of

criminal activity is a function of the proportion of the exogenous income not

reported.

 

Both the probability and the severity of punishment are found to deter crime

for a risk averse person. For risk lovers, the effect of the severity of punishment

is uncertain. An increase in the severity will have similar effects for illegal

activities as a wage decrease in labor supply models will have for legal

activities. Two effects obtain: a substitution effect and an income effect. The

substitution effect of a more severe punishment will consist in less crime. The

sign of the income effect will depend on individual attitude towards risk. For

a risk lover the income effect is positive, and the total effect on crime of a

change in severity becomes indeterminate. The effects of changes in gains from

crime and in exogenous income depend on whether there is decrease or

increase in the risk aversion or risk preference. For the common assumption of

decreasing absolute risk aversion an individual will allocate a larger proportion

of his income to tax cheating the higher his exogenous income and the higher

the gains from crime.

 

Heineke (1978) has presented a somewhat different type of model where the

individual allocates his time (and not his wealth or income) between legal and

illegal activities. The individual’s income is assumed to be equal to the sum of

three elements: exogenous income, the monetary and monetized benefits and

costs of legal activities, and the monetary and monetized benefits and costs of

illegal activities. (Monetization implicitly takes place if an individual, having

to choose between actions involving non-monetary gains and losses, acts

rationally according to certain axioms.) If convicted, this income is reduced by

a factor that represents the monetary and monetized costs of crime. Here, some

of the individuals may choose to specialize in either legal or illegal activities,

whereas others may choose a mix of the two. A marginal increase in the

probability or the severity of sanctions will affect the optimal mix of activities,

whereas such an increase may be insufficient to have an effect on individuals

who have specialized in one of the two activities. Assuming leisure time not to

be fixed, the same comparative statics results as for the portfolio choice model

are obtained. The reason for this similarity is the monetization of psychic

benefits, and the high degree of independence between the types of activities.

In addition, for some attitudes towards risk, it turns out that an increase in

returns to legal activity increases time allocated to both types of activities.

 

Several authors, first and foremost Ehrlich (1973), have studied the latter

type of model, but with the additional restriction that time allocated to leisure

is fixed (and thus independent of returns and costs for legal and illegal

activities). The assumption of a fixed leisure time obviously requires that the

time allocated to legal and illegal activity changes in opposite direction (and

with equal amounts), but the effects of changes in some of the parameters are

also different from the previous model. Whereas the effects on crime of changes

in exogenous income and gains to crime are the same as above, the effects of

changes in the severity of sanctions become inconclusive without further

restrictions on some parameters.

 

The portfolio model of time allocation with non-fixed leisure time has been

somewhat extended by Wolpin (1978) and by Schmidt and Witte (1984), who

have introduced four possible criminal justice states, each taking place with a

certain probability. In these models the effects of changes in sanctions, and in

gains and losses of crime become more ambiguous than in the previous models.

Especially, and somewhat surprisingly, illegal activity will decrease with

increasing unemployment under the standard assumption of decreasing absolute

risk aversion. The explanation is that unemployment implies a lower income,

and therefore a higher risk aversion, and then again a lower expected utility of

crime. Under risk neutrality time allocated to illegal activity is not affected by

a change in the expected employment rate. Baldry (1974) introduces the

assumption that a person has to choose between zero or a given number of

hours of legal work per week. Transforming the Ehrlich model into a nonlinear

programming model, he obtains unambiguous predictions of the effects on

crime of changes in sanctions and economic variables.

 

If one is not willing to accept the assumption that all psychic factors

associated with legal and illegal activities can be monetized, one has to use

utility functions where time allocations and their attributes are introduced

explicitly. Block and Heineke (1975a) have studied a model where a vector of

attributes of the penalty, interpreted as the length of sentence, is included in the

utility function. In this model one obtains considerably more ambiguous results

than for the previous models. Unless one is willing to make strong assumptions

about individual preferences, it is not possible to decide whether criminal

activity will decrease or increase as a result of changes in the probability of

punishment, of changes in returns to legal and to illegal activity, and of

changes in exogenous income.

 

Block and Heineke (1975a) have shown that changes in legal and illegal

remuneration lead to changes in illegal activity that are composed of stochastic

counterparts of the substitution and income effects of traditional supply and

demand theory. But the similarity is not close. Even if one assumes that illegal

activity is inferior (that is, that such activity is decreasing with income), it is

not possible to sign the relevant terms. Increasing the penalty, for instance, will

not unambiguously deter crime.

 

Witte (1980, p. 59) and Schmidt and Witte (1984) have studied a simplified

version of their several sanctions model where time spent in legal

income-generating activity (work), time spent in illegal income-generating

activity (theft, and so on), time spent in legal consumption activities, and time

spent in illegal consumption activities (drug use, assaultive activities, and so

on) are separate arguments in the individual’s utility function. Here too, similar

inconclusive results are obtained. When benefits and costs of legal activities are

risky, even more ambiguous results are obtained.

 

The standard assumption that people maximize expected utility is appealing

because it follows from the von Neumann-Morgenstern axioms of individual

behavior that many scholars regard as reasonable, or at least as a fruitful

hypothesis. However, many laboratory experiments have shown that people do

not always choose in accordance with these axioms, in particular Lattimore,

Baker and Witte (1992), who included burglaries in a set of risky prospects to

choose between. As a result, various alternative forms of preference functions

that are non-linear in the probabilities have been proposed. Eide (1995) has

substituted the assumption of rank-dependent expected utility for the ordinary

expected utility in various models of criminal behavior. In the latter study it is

shown that the qualitative results of comparative statics analyses are the same

for both types of assumptions.

 

Summing up the comparative static results so far, an increase in the

probability of clear-up or arrest has, regardless of the sign of the attitude

towards risk, a negative effect on the supply of crime. The effect of an increase

in the conviction rate, given arrest, is indeterminate without further

assumptions, and the same holds true for an increase in the probability of

imprisonment given conviction. However, reasonable assumptions will produce

the same conclusions as for the unconditional probability of arrest or

conviction. These results constitute a certain support for the probability part of

the deterrence hypothesis.

 

For any attitude towards risk in Beckers’s model, an increase in the severity

of punishment has a negative effect on the supply of crime. For the group of

portfolio choice models as a whole the severity part of the deterrence hypothesis

hinges upon the question of attitude towards risk. The effect of more severe

sanctions is especially uncertain for risk lovers, whereas risk averters in most

models offend less when sanctions increase. Furthermore, a positive shift in

punishment in the several-sanctions model and in the labor supply model with

non-monetized attributes can cause an increase in crime for any attitude

towards risk. In the latter model, the restriction necessary to generate this effect

is that the income effects must be greater than the substitution effects. The labor

supply models with non-monetized attributes give inconclusive effects also for

changes in the other parameters that are studied. For the other models the

effects of changes in the gains to crime, in exogenous income, and in income

from legal activities depend on the individual’s attitude towards risk.

 

As a whole, one may conclude that the effects of changes in the

environment depend on the individual’s attitude towards risk. If one is willing

to stick to the rather common assumption of decreasing absolute risk aversion,

and also that psychic effects can be monetized, and that there is just one type

of sanctions, the effects are clear: crime is deterred by increases in the

probability and in the severity of punishment, and enhanced by increases in

exogenous income, and in gains from both legal and illegal activities. The

reason why increases in various incomes and gains increases crime is that

punishment in the case of decreasing absolute risk aversion produces a smaller

reduction in expected (total) income. For risk-neutral people an increase in the

probability or severity of punishment and a decrease in the gains to crime will

reduce the supply of crime, whereas changes in exogenous income, and in the

remuneration of legal activity have no effect. Here, changes in the latter income

components do not change the bite of punishment.

 

A crucial assumption in the studies mentioned above seems to be the

Bernoulli distribution of the probability of punishment. Introducing a more

general distribution of risk into the Becker-type model of Block and Lind

(1975a), Baldry (1980) concludes that the ‘standard’ deterrence results cannot

be derived.

 

A good survey of the main contributions to the development of the economic

models of crime is found in Schmidt and Witte (1984).

 

4. The Benefits and Costs of Crime

 

Various studies have elaborated on the benefits and costs of crime. The gains

and losses included in the economic models of criminal behavior are usually

meant to represent all kinds of benefits and costs that have an effect on the

people’s decisions. People are assumed to allocate time to criminal activity until

marginal benefits equal marginal costs. For some people marginal benefits are

probably always lower than marginal costs, and we then have a law-abiding

person. Others will specialize in crime, whereas most of us possibly commit an

offense now and then.

 

The kinds of gains obtained from a criminal act vary, depending on the type

of crime and the individual criminal: some are monetary, obtained from theft,

robbery, insurance fraud, and so on. Others are psychic, such as the thrill of

danger, peer approval, retribution (bank robbery), sense of accomplishment, or

‘pure’ satisfaction of wants (rape). For some property crimes the prices

obtained on markets of stolen goods are of importance.

 

Among the costs one may distinguish between material costs (equipment,

guns, vehicles), psychic costs (guilt, anxiety, fear, dislike of risk), expected

punishment costs and opportunity costs.

 

The punishment costs include all formal and informal sanctions, as well as

pecuniary costs arising from lawsuits (lost income and lawyer’s fee). The

formal sanctions include fines, various forms of incarceration, and so on. The

more severe these sanctions are, the higher the cost. The informal sanctions

include any personal inconveniences connected with arrest, suit and conviction.

The sanctions related to the social stigma caused by arrest and formal sanctions

must be added. The nuisance associated with appearing in court, and the

reactions of employer, family and friends, might have a stronger effect than

formal sanctions.

 

The opportunity cost of crime consists of the net benefit (gross benefit

minus cost) of the legal activity forgone while planning, performing and

concealing the criminal act. The lower an individual’s level of income, the

lower is his opportunity cost of engaging in illegal activity.

 

The amount a person can earn in the legal sector may depend upon age, sex,

race, education, training, region, rate of unemployment, IQ, and so on. People

able to earn only a rather low wage will have a low opportunity cost of crime,

the cost of giving up legal income. We would therefore expect that among

criminals there are more young people, men, blacks, low-paid workers, and so

on than in the population at large. This is in fact what crime statistics tell us,

but more refined empirical studies are necessary to substantiate such

relationships.

 

Many individual characteristics might have an effect on benefits and costs.

Individual rates of discount might be important. The gains from crime often

occur immediately, whereas punishment is something that might come in the

future, and stretched over a long period of time. A high discount rate will

therefore tend to increase crime. The probability of punishment will be different

for different people. Some are more clever than others at concealing the offense

and eluding the police. There are also differences in abilities of defending

oneself in court, or in engaging good lawyers. The attitude towards risk will

also have an effect.

 

A high rate of recidivism is in accordance with the model of rational choice.

If for an offender preferences are stable and the opportunities available remain

the same, the degree of criminal activity will not tend to decrease after a

conviction. Recidivism is thus not necessarily a result of erratic behavior or

lack of self-control, but rather a result of rational choice. Moreover, several

factors that count in favor of crime are increased by imprisonment: additional

criminal skills are acquired, and opportunities of legal income are reduced. If

it was rational to commit a crime in the first place, it is all the more so after

having served a prison sentence. If the sentence has increased the criminal’s

evaluation of how probable or severe sanctions might be, or if he or she has

obtained some benefit from prison education schemes, the tendency to

recidivism will be counterbalanced.

 

5. Are Criminals Really Rational?

 

Exaggerating somewhat the differences between sociologists and economists,

one may say that the first consider crime as deviant behavior whereas the latter

consider it as rational. Figure 1 illustrates the main elements characterizing an

individual’s choice situation according to the theory of rational choice. The

individual has a feasible set of courses of action, some of which are illegal. The

environment, including sanctions and wages, determines the outcomes of the

various courses of action. The individual is assumed to choose the course of

action that best satisfies its preferences. Preferences include not only wants, but

also norms. The guilt of acting in conflict with norms is part of the costs of

crime.

 

 

 

 

 

In theories of economics of crime, norms are seldom studied, or even

mentioned. Preferences as a whole are usually assumed to be constant, and

authors do not find it necessary, or do not feel competent to discuss norms.

Traditional criminological theories, on the other hand, suggest that the

individual’s environment has a significant impact on people’s preferences,

especially on norms, but also on wants. Theories about culture conflict, cultural

deviance, anomie and learning relate individual preferences to various

characteristics of the society. Other theories suggest that preferences are

inherited or dependent on age, gender, race, intelligence and other personal

characteristics. In the literature of economics of crime these various theories are

often neglected, and the main question studied is how the environment

produces incentives to commit, or not to commit, crimes.

 

In a very broad sense a deterrent is any factor that exerts a preventive force

against crime. Sanctions may have an effect on crime either by causing fear or

by influencing norms. The combination of these effects is in parts of the

literature on crime called ‘general prevention’ (Andenaes, 1975). In economics

of crime one focuses on the effects of law enforcement on the outcomes of

actions, and thereby on illegal behavior. This is the deterrence mechanism in

the narrow sense. The possibility that law enforcement or other aspects of the

environment might affect individual norms and wants, for example, by

conditioned aversion as suggested by the behavioral perspective, is given less

attention.

 

An interesting question is whether the model of rational choice is in conflict

with, a substitute for, a supplement to, or a general framework for other

theories of crime. Carr-Hill and Stern (1979) emphasize that the economic and

criminological approaches should be seen as complementary rather than

conflicting. They maintain that the economic approach isolates the importance

of the probabilities and magnitude of reward and punishment, and shows how

they can be treated formally. The criminological approach takes these for

granted and indicates how different groups might view and react to these

probabilities, rewards, and punishments.

 

These two approaches are related to the issue of opportunity vs. motivation

as explanation of crime. Economists and others who focus on costs and benefits

of crime in a rational choice framework, also take into account that crime

presupposes potential victims. The better the opportunities of hitting valuable

and low-risk targets, the more crime there is. Those who more or less explicitly

dismiss the theory of rational choice often focus on the motivation of

individuals, assuming that behavior is determined by individual characteristics

and by the norms of the groups to which they belong.

 

The opportunity approach is an element in the market models of crime,

where the number of offenses is determined by the interaction of potential

offenders, who are seeking the best targets, and potential victims, who by

measures of private protection seek to be less attractive or vulnerable to crime

(compare Ehrlich, 1981, 1982, 1996, and Cook, 1986).

 

Several authors have discussed whether people have sufficient information

about the environment and about outcomes of actions to make rational choices.

Becker and others maintain that even if choices are based on subjective beliefs

that are wrong, the choices are meaningful from a subjective point of view, and

behavior can be explained and understood on this basis. One may argue that

this is not a satisfactory answer to the claim that people have cognitive

limitations, and that they stick to ‘satisficing’ and not to maximization. The

studies of Caroll and Weaver (1986), Tunnell (1992), and Nagin and

Paternoster (1993) suggest that Simon’s theory of bounded rationality might be

a better representation of offenders’ behavior than the rational choice theory,

a conclusion that is supported by Niggli (1994).

 

It has also been argued that the simple rational choice theory is inadequate

because people’s behavior is determined by procedural rationality, in which an

individual is portrayed as a follower of rules established by history or social

relations, or by expressive rationality, in which an individual, through symbolic

acts, demonstrates to himself and others his self-conception and worth. There

is disagreement about how serious such criticism is for the use of the rational

choice theory in studies of crime. Ehrlich (1973, p. 532) maintains that ‘[s]ince

those who hate need not respond to incentives any differently than those who

love or are indifferent to the well-being of others, the analysis ... would apply

... to crimes against the person as well as to crime involving material gains’.

 

6. Empirical Studies

 

In a great number of empirical studies the theoretical models of criminal

behavior have been tested, and the effect on crime of the probability and

severity of punishment, and of benefits and costs of legal and illegal activities

has been estimated. The influence of norms, tastes, and abilities, corresponding

to constitutional and acquired individual characteristics, have in some cases

been studied indirectly by including variables like age, race, gender, and so on.

A variety of equation specifications and estimation techniques have been used,

and the studies have been based on data from countries and states down to

municipalities, campuses, and individuals.

 

Analogously to the terms psychometric studies, cliometric studies, and so

on it seems appropriate to introduce the term ‘criminometric studies’ to

characterize this field of research. The subject matter is crime, and it gives the

field a somewhat distorted and too limited range to call these studies

econometric, although this is what is usually done. The studies are rooted in a

general theory of rational choice, and not in some rational choice theory

presumably limited to economics.

 

In the framework of norm-guided rational behavior norms may depend on

the environment. In most criminometric studies norms, as well as wants, are

assumed to be constant, and often also equal among individuals. Becker (1976,

p. 5) expresses a rather common attitude by stating that ‘[s]ince economists

generally have little to contribute, especially in recent times, to the

understanding of how preferences are formed, preferences are assumed not to

be very different between wealthy and poor persons, or even between persons

in different societies and cultures’. With this assumption it is relatively easy to

test other parts of the theory, such as hypotheses about the effect of sanctions,

and of gains and losses of legal and illegal activities. If preferences differ

among individuals, estimates of the effects of sanctions will be relevant for an

‘average’ person. The explicit assumption that individual preferences are

constant, distinguish criminometric studies from most other studies in

criminology.

 

There are good reasons to carry out empirical studies of criminal behavior

at the individual level instead of an aggregated level. In the first place it is at

best controversial to posit that behavior is anything but individual. Second, the

theoretical models that are developed are based on individual rational choice.

Third, as will be discussed below, studies based on aggregated data require a

number of additional assumptions of questionable validity. Fourth, the

statistical identification problem is less serious when individual behavior is

studied. Using aggregated data one faces the problem of distinguishing between

the effect of the probability of arrest on the amount of crime and the effect of

the amount of crime on the probability of arrest. In empirical studies at the

individual level it can reasonably be assumed that the probability and severity

of punishment is determined without being influenced by the actions of a given

individual. Thus, the deterrence variables can be considered to be exogenous

to the individual’s choices, and the problem of simultaneity inherent in macro

studies is absent. Unfortunately, empirical tests of these models by use of

information on individuals are few. The application of the theoretical models

to empirical studies is intricate (Manski, 1978), and suitable data are scarce.

The data we have are mainly self-reports on criminal activity, and records of

criminal activity compiled by the criminal justice system. The most serious

problem with the latter type of data is that they do not constitute representative

samples of the population, but are biased in the sense that only convicted

persons are included. It is hardly possible to test a general theory of rational

criminal behavior by studying only one subgroup of the offenders. A related

problem is that most available data sources include information only about

choices made, and not about those available, but not chosen. It is difficult, if not

impossible, to test a theory of rational choice if the choice set in this way is

limited. Whereas such data are of limited interest for studies of general

deterrence, that is, of effects on people in general, they are useful for studies of

special deterrence, that is, of effects on the individuals that are punished.

 

The bulk of criminometric studies consists of cross-section regression

analyses based on macro data. Some of them are rather broad, including many

types of regional areas, estimation techniques and types of crime, whereas

others concentrate on particular types of crime, such as property crimes or

hijacking. A few of them address special questions, such as the effect of police

‘aggressiveness’ in patrolling, or the influence of income differentials. Time

series studies are less numerous, and employ mostly data on total crime.

 

The majority of these empirical studies of crime have been evaluated in

various surveys. In an annotated bibliography Beyleveld (1980) reviews a

number of investigations of correlations between crime and deterrence

variables, in addition to 35 cross-sectional and 31 time-series econometric

studies of crime. A thorough review of the empirical evidence of general

deterrence is also given by Nagin (1978), who comments on 24 correlation and

econometric studies, all but two also covered by Beyleveld. Taylor (1978)

concentrates upon six major econometric studies, whereas Pyle (1983) reviews

the same studies and about 15 others. A shorter, but somewhat more up to date

review is Cameron (1988), and a more recent comprehensive survey is found

in Eide (1994). Reviews of certain parts of the literature are found in Passell

and Taylor (1977), Fisher and Nagin (1978), Klein, Forst and Filatov (1978),

Nagin (1978), Vandaele (1978).

 

6.1 Empirical Effects of Punishment Variables

In empirical studies the measures used to represent the probability of

punishment include the probabilities of arrest, of clearance, of conviction, and

of conviction given arrest. The severity of punishment is represented by fines,

by the length of sentence, or by time served. Witte (1980) and Schmidt and

Witte (1984) have employed individual data on post-release activities of a

random sample of 641 men released from prison in North Carolina. The effects

on crime of measures of both the probability and the severity of punishment are

found to be more or less negative. Myers (1983), using a sample of 2127

individuals released from US Federal prisons, finds that severity of punishment

has a statistically significant negative effect on crime, whereas the probability

measure (the ratio of previous prison commitments to previous convictions) has

a positive effect. Higher wages are found to reduce recidivism. Trumbull (1989)

has used data on about 2000 offenders released from prisons in North Carolina

to study recidivism and special deterrence. He finds that none of the deterrence

variables (probabilities of arrest, conviction and imprisonment, and length of

sentence) are statistically significant. Trumbull finds this result natural, since

the sample consists only of individuals who, whatever the probability and

severity of punishment, have chosen to engage in illegitimate activities.

However, an increase in an offender’s own previous sentence length has a

significant negative effect on crime, a result that corroborates the hypothesis of

special deterrence. Higher earnings on the first job after release has a negative

effect on crime. Quite unexpectedly, so has unemployment. Viscusi (1986b)

uses an approach common in labor economics in the studies of hazardous jobs

to estimate the risk/reward trade-off for illegal activities. In labor markets

increasing health risks are often rewarded by some amounts of money in

addition to non-risk wages. Treating the probability and severity of punishment

in the same manner as the probability and severity of injury are treated in

analyses of hazardous jobs, Viscusi is able to estimate the effects of changes in

these variables. A survey of 2358 inner-city minority youths from Boston,

Chicago and Philadelphia constitutes the data employed. Viscusi finds that the

premiums obtained for criminal risks are strong and quite robust. In his

framework this is interpreted as a corroboration of the general deterrence

hypothesis.






中国法律经济学网登载此文出于学术研究之目的,绝不意味着中国法律经济学网赞同其观点或证实其描述。以上内容仅供研究者学习与交流,无意侵犯版权。如有侵犯您的利益,请告知。我们将尽快删除。

加入日期:2007/1/11 9:08:11浏览次数:2522
发表评论
名号:
内容:
验证: 2576
法律经济学网
联系站长: 柯华庆 lawgame@263.net 京ICP备09028584号
北京市昌平区中国政法大学法学院(102249)
本网站由卡卡鱼网提供技术支持 网站总访问量:1996356