??????????? ??? ???????? ???? ?????????? ???? ????? ??????? ????? ??? ?????? ???????????? ????????????
Communities that cannot provide an acceptable level of security for persons and
property will not long survive. Economists have long pointed out that even if such
communities were to survive they would not prosper. Communities where the strong are
allowed to freely victimize the weak will be small and surviving members will have to
spend most of their resources on defense.
When property can be freely taken by theft and deception, no one has the
incentive to invest. Protection of property from taking is the most basic of all property
rights. Without this protection, the problem of the common is pervasive.
Adam Smith believed that the protection of person and property was the most
important duty of government after national defense. Yet, somewhat surprisingly
economists interested in public economics have rarely analyzed the nature of the
government’s roll in providing domestic security for citizens. A search of both general
and advanced textbooks on public economics revealed no text that considered
government expenditures and the government roll in crime prevention and criminal
In this paper, we seek to use the perspective and tools of public economics to
examine crime control and criminal justice. We begin by presenting both general and
specific measures of the level and nature of crime for a variety of countries.
Unsurprisingly crime is pervasive. However, the level of crime varies substantially across
countries. In the section that follows, we outline the arguments for at least some public
provision of crime prevention, enforcement, prosecution and defense, and adjudication.
We briefly consider sentencing. In Section 4, we describe the relative rolls of the private
and public sectors in the provision of crime control and criminal justice. In the
penultimate section, we summarize some research on the effectiveness of public
expenditures on crime control and criminal justice. We conclude by suggesting some
potentially productive research directions.
II. WHAT IS THE PROBLEM?
Crime is everywhere. Even though countries define crime differently in their
criminal codes no country is without crime. Some things are crimes almost everywhere.
These have come to be called the core or traditional crimes—murder, robbery, rape, theft,
burglary, fraud and assault. Even for these crimes, measuring the extent of the activity is,
to say the least, difficult. Perpetrators have strong incentives to keep their activities secret
and discovery by public or private enforcement agencies is limited both by resources and
Crime is, of course, not limited to the traditional crimes. Trafficking in illegal or
stolen property is widespread and increasingly transnational. Financial and environmental
crimes, sometimes carried out by large multinational enterprises, can and often do cause
mass human suffering and financial losses. One need only think of the Bhopal disaster or
the BCCI scandal.
Obtaining an overall measure of the extent of crime that is comparable across
countries is a daunting task. Fortunately, researchers at the World Bank have carefully
compiled, analyzed and aggregated indicators for the “rule of law”, graft and political
instability and violence from 13 different sources for over 150 countries. See Kaufmann,
Kraay and Zoido-Lobaton (1999a, 1999b) for a list of sources and methodology used.
They provide an aggregate measure for each indicator that ranges from –2.5 to 2.5.
Countries with higher numbers are deemed to be more law abiding. Kaufmann, et al.
provide standard errors as well as point estimates for each country.
Kaufmann’s, et al. rule of law provides, as far as we are aware, the broadest (in
terms both of types of crime and geography) indicator of crime that is available. To
obtain a measure for the rule of law, Kaufmann, et al. aggregate a number of indicators
that measure the extent to which agents have confidence in and abide by the rules of
society. The indicators include perceptions of the incidence of both violent and
nonviolent crime, the effectiveness and predictability of the judiciary and the
enforceability of contracts. Kaufmann, et al. indicate that together these indicators
measure the success of a society in developing an environment in which fair and
predictable rules form the basis for economic and social interaction.
Figure 1 displays Kaufmann, et al. measures for the rule of law for selected
countries. The diamond in the center of the country name is the point estimate. The lines
emanating from the point estimates provide the 90% confidence interval for each
Insert Figure 1 here
From these data it is clear that some countries (e.g., Switzerland, Singapore) have
much stronger rules of law than other countries (e.g., the Congo, Iraq). It is also clear that
the range of reasonable estimates for the middle range countries (e.g., Brazil, India,
Turkey) do not differ significantly. Still, these estimates provide a useful broad
assessment of the extent of the crime problem across countries.
Kaufmann ‘s, et al. indicator for graft is designed to measure perceptions of
corruption, an important aspect of crime (see Figure 2). Kaufmann, et al. describe this
measure as indicating the degree to which public power is used for private gain. This
measure of crime is narrower than the rule of law considered previously, but still it
provides a valuable measure of the extent of an important and often overlooked aspect of
crime. Rankings are generally similar, but not identical to those for the rule of law. For
example, Italy is quite like Spain in terms of the rule of law, but has a much lower rating
than Spain for graft. Tunisia is quite like Brazil in terms of graft, but Tunisia has a higher
measure for the rule of law than does Brazil.
Insert Figure 2 here
While graft measure public property offences, Kaufmann’s, et al. political
instability and violence might be considered a measure of the likelihood of violent crime
related to government. Figure 3 present measures public perceptions regarding the
likelihood that the government in power will be destabilized or overthrown by
unconstitutional or violent means. On this indicator Italy ranks above Spain and Algeria
joins the Congo and Iraq at the bottom of the scale.
Insert Figure 3 here
Turning from general measures of lawfulness to measures for the core crimes, one
finds less information both in terms of the number of countries for which comparable
data are available and the crimes for which the extent is measured. Before proceeding, it
is important to note that reports of crime to the police, the most broadly cited statistics in
many countries, are generally more a measure of the functioning of the criminal justice
system than of crime (Newman, 1999).
That being said some crimes are more difficult to hide than others. For example,
murder is generally known because a dead body rarely remains successfully hidden for
long. In stable countries, the number of murders (homicides) is generally quite well
recorded by the police.1 Further vital statistics provide a check on the number of
homicides in countries with well-functioning public health departments.
The United Nations Crime and Justice Survey (UNCJS), the International Police
Organization (INTERPOL) and the World Health Organization (WHO) provide
independent estimates of the extent of homicide for a variety of countries. WHO reports
only on successful, completed acts of homicide from vital statistics. UNCJS and
INTERPOL rely either directly or indirectly on the reports of national criminal justice
As can be seen in Figure 4, the three sources of data generally agree rather closely
on the homicide rate for most countries. However, there are notable exceptions. For
example, the different sources provide quite different estimates for Bulgaria and the
Russian Federation. While the sources of data indicate quite different magnitudes for
these countries, all sources indicate that these countries have higher murder rates than
other countries for which data were available. For other countries, homicide rates range
from under 1 per 100,000 (i.e., Japan and Norway) to over 2 per 100,000 (e.g., Finland
1 Newman (1999), p. 11.
2 Raw data on homicide from the UNCJS and INTERPOL include attempts as well as completed murders. We use figures given in the
notes to Box 0.7 of Newman 1999.
For crimes other than murder measurement is more difficult. Surveys of victims
provide reasonably accurate measures of offenses for which there is an identifiable victim
that knows that he/she has been victimized and is willing to report it in a survey setting.
For international comparisons, the International Crime Victim Surveys (ICVS) that were
carried out in 1989, 1992/1994 and 1996 provide results for a number of industrialized,
transition and developing countries.3 These surveys use consistent definitions of offenses
while official statistics depend on the definition of offenses in the criminal law, which
can vary widely from country to country. The surveys asked about the following
offenses: (1) contact crimes (robbery, sexual offences, threats and assaults), (2) burglary
(including attempts), (3) car crimes (car theft, theft from car and car damage) and (4)
other thefts (motorcycle theft, bicycle theft and other personal theft).
These surveys reveal that more than half of urban residents report having been a
victim of one or more of the covered offenses during the last five years. Being the victim
of a crime is a common occurrence in all urban areas. Rates of victimization are highest
3 For a description of methodologies and countries included, see Note 3 to Box 0.9 of Newman (1999).
Homicide Rates per 100,000 Population in 1993
Rep. of Korea Russian Fed.
Rate per 100,000 Population
Source: Notes to Box 0.7 of Newman (1999)
in Africa and Latin American where almost ¾ of urban residents report having been
victimized during the last five years. Rates are lowest in Asia where 45% of urban
residents report victimization over a five-year period. Victimization rates in the US,
England and Wales, and West Germany are quite similar with between 60% and 65% of
urban residents reporting victimization during a five-year period.
The ICVS asked respondents about their perception of the relative seriousness of
various types of crimes. Western European and North American countries tended to rank
violent crimes (e.g., robbery with a weapon) most seriously while African, Asian, Central
and Eastern European, and Latin American countries ranked car theft as the most serious
of the offense considered.
Urban residents have over a 50% chance of being the victim of a contact crime
during a five year period in Columbia and less than a 10% chance of being the victim of a
contact crime in Switzerland. In Western Europe, the Netherlands reports the highest
level of victimization for contact crimes (22% of urban residents report victimizations).
This reported rate of victimization is approximately the same as Russia’s. In the US,
approximately 20% of urban residents report that they have been victims of contact
crimes during a five-year period.
Car theft is most frequently reported in New Zealand (just under 50% of urban
residents report car thefts during a five year period) and least frequently reported in China
(less than 2% of the Chinese report car thefts during a five year period). These numbers
point up the importance of opportunity. For most property crimes, victimization rates are
higher in wealthier countries where there is more of value to steal. Theft is significantly
correlated with holdings of durable goods.
It is more difficult to gage the extent of crimes other than murder and those
covered by the ICVS. Official records of crime such at the US Federal Bureau of
Investigations (FBI) Uniform Crime Reports (UCR) reflect many things in addition to the
underlying crime rate (e.g., the willingness of residents to report crimes to the police, the
reporting practices of police agencies). We discuss a few attempts (mainly cross
national) to study other types of offending here.
The ICVS like most victimization surveys is concerned with offense where
individuals or households are the victims. A few victimization surveys have also sought
to discern the extent to which businesses are victimized. For example, a 1993 survey of
commercial establishments in England and Wales reports that 8 out of 10 retailers and 2
out of 3 of manufacturers experience one or more crimes covered by the survey in 1993.
Commercial victimization appears to be highly concentrated with 3% of retailers
experiencing 59% of the crime reported in the survey. The reported risk of victimization
and the amount of the loss was higher for retailers and manufacturers than for households
(Mirrlees-Black and Ross, 1995a, 1995b).
As is well known, the core crimes with the exception of fraud is like basketball
primarily a young man’s game. In a very interesting study, Junger-Tas, Gert-Jan and
Klein (1994) report the results of surveys of young people (14-21) in 12 countries that
were carried out in 1992. As can be seen in Figure 5, rates of self-reported offending vary
substantially across the areas studied. For example, young people in Athens report the
highest rates of violent crime with more than half of the respondents reporting that they
had committed an offense during the last year. Young people in Helsinki report the
highest rates of property crime (just under 50%) and young people in England and Wales
the highest rate of drug offending (26%).
Regardless of whether one considers general measures of law abidingness,
victimization reports or self reports of crime by the young, the message is the same---
there is a lot of crime. That being the case we turn to the next issue. What, if anything
should the public sector do about it?
III. WHAT PUBIC ROLL?
As H.L.A. Hart has noted, communities must have restrictions on the free use of
violence, theft and deception if they are to survive and prosper (Hart, 1994, p. 91). To put
it somewhat differently, societies will place restrictions on the core or traditional crimes,
murder, rape, robbery, theft, burglary, fraud and assault.
The question is how these restrictions will be imposed. In what Hart calls
primitive societies, these restrictions may be imposed by custom and informal means of
Percent Reporting Offenses During Last 12 Months
England & Wales
3 Itlanian Cities
Source: Box 0.10 of Newman (1999)
social control. However, as communities grow, develop and become more heterogeneous,
restrictions against committing the core crimes tend to become a matter of religious or
governmental rules or a mixture of both.
Today, the legal systems that outlaw the core crimes can be classified broadly into
three types: (1) civil law, (2) common law and (3) Islamic Law (Newman, 1999). Only
in Islamic law do we find a mixture of religious and governmental restrictions against the
core crimes. Islamic law tends to encourage non-governmental response to the core
crimes, including murder (Groves, Newman and Corrado, 1987). Islamic law tempers
retaliation by encouraging forgiveness. Under Islamic law, a victim or a victim’s family
may waive retaliation and receive instead a money payment from the perpetrator of the
crime. In such a situation the government will not become involved with either the
offender or the offense. Only if the perpetrator and victim cannot agree does the
government become involved.
In both the common law and civil law traditions, it is the government who forbids
the core crimes. The civil law tradition originated as a combination of Roman Law and
papal statements of the Roman Catholic church. However, most countries that follow the
civil law tradition today (e.g., France, Germany, Brazil, Russia, Indonesia) rely on
secular legislation as the source of restrictions against the core crimes. Under a civil law
system, there is a sharp separation of powers. Legislatures make the law and judges apply
By way of contrast, the common law tradition relies on the customs of the people
as its original source. Much of common law became codified over time. For example,
restrictions against the core crimes are now generally legislated. However, judges may
and do interpret the legislated law in particular cases. The judge has the ability to make
laws in the common law tradition, but not in the civil law tradition. Countries that use the
common law tradition include England, India and the United States.
Why do most developed countries empower one or more arms of government to
determine what is criminal? The rise of the power of government to determine what is a
crime coincides with the rise of the nation state, industrialization and urbanization. The
rulers of the new nation states sought to monopolize control over the use of force in the
4 In developed countries the Civil Law and Common Law traditions have become more similar during the post World War II period.
hands of their servants for quite obvious reasons. Industrialization and urbanization
weakened the force of traditional or customary law. The rise of secular society weakened
religious restrictions against the core crimes although religious restriction can still be
very important in preventing crime.
Standardization and codification of criminal laws lowered the transaction costs
required for both inter-personal interaction and property transfers. Reliable and
standardized rules allowed for greater economies of scale by lowering the costs of trade
and travel over larger areas. As noted earlier, protection of property from taking by theft
or fraud is the most basic of all property rights and necessary to prevent the problem of
the commons. The newly emergent nation state was in a unique position to undertake the
codification and had the incentive to do so to solidify its power.
Many believe that crime is very distinct from other types of illegal acts because
crimes are believed not only to harm the victim, but also to harms other member of
society. Crimes, particularly the core crimes, are believed to be offenses against society
not merely offenses against the individual (Kaplan, et al., 1991; Cooter and Ulen, 2000).
In economic terms, crime produces negative externalities and will be too high if there is
no public intervention.
An additional distinguishing feature of crimes as opposed to most civil offenses,
such as torts, is that conviction of a crime requires a finding of mens rea (a guilty mind).
If enforcement costs were zero, we might want to eliminate the core crimes. By way of
contrast, we only wish to encourage and efficient level of prevention expenditures in the
case of torts. The intent to do physical harm or to transfer property by force or deception
requires stronger deterrence than do accidental occurrence and may even justify putting
the offender in a position where he can no longer offend (i.e., incapacitating the
Some have argued that many crimes should be handled like torts with the victim
bringing a case for compensation (e.g., Friedman, 2000). Others argue that crimes are
very different from torts. For some crimes (e.g., rape, robbery), some argue that no
amount of compensation can make the victim whole (Cooter and Ulen, 2000). Others
argue that the intent required for a crime requires punishment/retaliation (United States v.
Regardless of the position one takes in these arguments, even the most ardent
proponents of treating many things that are currently crimes as torts find a need to
criminalize in some circumstances. For example, Friedman suggests that criminalization
is necessary for acts in which the victims are anonymous (e.g., highway robbery) or
defendants are “judgment proof” (too poor to pay compensation to victims). The matter
in dispute is what acts should be criminal.
Police do many things, but we are concerned only with their crime control
functions. These functions include crime prevention and the apprehension of offenders.
(a) Preventing Crime
What is the proper roll of government in preventing crime? As we will show
later, crime prevention is now generally split between the private and public sectors with
the private sector generally spending more on prevention than the public sector.
Much crime prevention does not involve the use of force and, hence, the argument
for public provision is greatly weakened. When crime prevention does require the use of
force, the case for public provision is stronger (Hart, et al., 1997). Public monopoly of
and control over the legitimate use of force is central to the provision of standard levels
of security for person and property. Further, to be effective the police require the
sympathy and cooperation of the public. Public police may be better able to nurture these
Public police forces are relatively new and emerged along with large,
heterogeneous urban areas. They arose in response to a public outcry for more order. The
first public police force was established in London in 1829 to provide a full-time day and
night patrol to prevent crime. Sir Robert Peel who was mainly responsible of establishing
the force under the Metropolitan Police Act insisted that political patronage be excluded
from appointments and promotions (Miller, 1977 as excepted in Kaplan, et al., 1991).
Boston established a public police force in 1837 to prevent violence between Protestants
and Catholics. New York City’s Municipal Police Act established a semi-military day
and night patrol force in 1845. The New York force was not insulated from political
patronage (Kaplan, et al. pp. 141-145).
(b) Apprehension of offenders
Police are not only asked to prevent crime, but also to apprehend offenders when
crime does occur. Apprehension often requires the use of force and, hence, a reasonably
good case can be made for public provision (Hart, et al., 1997). Also, in contrast with
many civil offenses such as automobile accidents, the identity of the perpetrator may be
unknown to the victim (Polinsky and Shavell, 1999). Even if the perpetrator’s identity is
known the victim may well not be willing to press a case.
As noted earlier, crime can produce very high negative externalities (e.g., fear of
going out when someone has recently been murdered in the neighborhood) and so both
prevention of crime and apprehension of criminals will be inadequately supplied by the
private sector. The distribution of prevention and apprehension is also likely to be much
higher in wealthier than in poorer neighborhoods if these activities are financed by the
To summarize, both efficiency and equity arguments call for public financing of
at least a minimal level of police services in larger communities. The wealthy will often
choose to purchase additional prevention to protect both their person and their property.
The fact that the police’s crime fighting activities can involve the use of force in a
wide array of situations suggests that public provision rather than contracting with the
private sector will likely be best for these activities (Hart, et al., 1997). Other police
functions (e.g., record keeping) may well be better contracted out than provided by the
The roll of the state in persecuting crime clearly depends upon whether or not
crime is seen mainly as an offense against the state or as an offense against the victim.
Countries as diverse as China, France and the USA see crime mainly as an offense
against the state or the community as a whole and public officials generally prosecute
criminal cases. However, the powers and type of official empowered to prosecute crimes
varies substantially across countries. In Continental Europe, the prosecutor is an
appointed career civil servant of the central government with a close relationship to the
court. Most prosecutors in the United States are locally elected officials with substantial
autonomy from both the judicial and executive branches of government (Kaplan, et al.,
In China, private citizens (generally the victim or the victim’s relatives) prosecute
less serious crimes while the state prosecutes more serious offenses (Newman, 1999, p.
40). From an economic point of view, this split can be justified because of the larger
negative externalities arising from more serious crimes.
Traditional English common law sees crime as an offense against the victim not
as an offense against some broader community. It was not until 1879 that England
created an Office of Public Prosecutions. The Director of this Office is a career civil
servant. The actual trial of cases is assigned to barristers in private practice designated as
Crown Counsel. For a discussion see Newman (1999, p. 132) or Kaplan, et al. (1991, pp.
One can easily argue for public expenditures to secure prosecution of crimes that
cause large negative externalities since private parties would be expected to bring too few
suits. Arguments for public provision of prosecution generally rest on the need to
maintain a high level of standards in the prosecution of serious criminal cases. As
Kaplan, et al. express it: “The prosecutor is also a representative of the government upon
whom the courts, and society, impose a standard of ethics which may transcend any
particular rule” (1991, p.311). Contracts for prosecutors would tend to be quite
incomplete. Privately contracted prosecutors would have strong incentives to lower
standards in order to lower costs (Hart, et al, 1997).
Most developed countries whether following a civil or common law legal
tradition provide public funding to allow indigent defendants to hire legal counsel in
serious criminal cases. The need for representation is probably stronger under the
adversarial common law tradition where the judge can only consider the evidence
brought before her than under the inquisitional civil tradition where the judge can
actively search for information.
Economic argument for the provision of defense counsel for the indigent rest
mainly on equity grounds although it would be possible to argue that it is inefficient to
imprison the innocent poor. Horizontal equity dictates that the equally situated should be
treated equally. In terms of the core crimes, equal situation might well be defined as
being equally innocent or equally guilty. Defendants without legal counsel will clearly
not be in a situation that is equivalent to the situation of those who are able to purchase
competent legal counsel.
The USA has a large, complicated and much criticized “system” for providing
defense counsel for the indigent. The sixth amendment to the US Constitution establishes
the right to counsel in Federal criminal prosecution. However, most of the core crimes are
prosecuted at the local level not at the Federal level. During the 1960s and 1970s, a series
of US Supreme Court cases (e.g., Gideon v. Wainwright, 1963, Argersinger v. Hamilin,
1972) established indigents’ right to counsel for all criminal prosecutions that carry a
sentence of imprisonment.
The Supreme Court did not indicate how state and local government were to
provide indigent defense counsel or what source of funds would be used to pay for
indigent defense. Currently, local governments (e.g., the Counties) are primarily
responsible for providing defense counsel for the indigent. The majority of funding for
indigent defense comes from local government although the state share of funding has
grown over the years (Smith and DeFrancis, 1996).
The system for providing indigent defense in the US varies markedly from place
to place. However, three basic methods are used:
• Assigned counsel programs appoint indigent counsel on a case-by-case
basis. The counselor is chosen from members of the local, private bar.
• Under contract attorney programs, the local or state government
contracts with individual private attorneys, private law firms or local bar
associations to provide indigent defense.
• Under public defender programs, salaried staff of fill-time or part-time
attorneys provide indigent defense. The public defenders may be
employed by state or local government or by non-profit corporation
contracted to provide indigent defense (Kaplan, et al, 1991, p. 351).
About three-fourths of inmates in State prisons and about half of those in Federal
prison in the US received publicly provided legal counsel for the offense for which they
were serving time (Smith and DeFrancis, 1996). The quality of publicly provided legal
counsel varies widely both across the US States and, in many states, across local
jurisdictions. In 1986, the cost per case of providing defense counsel ranged from a low
of $63 in Arkansas to a high of $540 in New Jersey (Kaplan, et al., 1991, p. 353).
Horizontal equity is not achieved by the current US system for indigent defense.
Further, as noted by Posner, criminal defendants have less access to the private market
for lawyers than do defendants in civil cases where damage awards are possible (Merritt
v. Faulkner, 1983).
The need for impartial adjudication of guilt in criminal cases is widely accepted.
In traditional societies, this adjudication was often by a body of chiefs or elders. In the
course of economic development and urbanization, criminal law and criminal procedure
became increasing complex and a widely respected, informal body to adjudicate more
difficult to construct and use. The increasing complexity and impersonality of
adjudication has led to the increasing use of trained personnel to referee the
trial/inquisition (e.g., enforce criminal procedure) and establish matters of law. Still,
adjudication is in many countries a task that is carried out jointly by lay citizens and
For example, in the US, a jury of peers (selected from voter registration lists or
other compilations of residents) determines the facts and the quilt or innocence of the
accused. The judge referees the proceedings (e.g., enforces proper procedure and the
admission of evidence), instructs the jury, determines matters of law and decides on the
sentence a convicted defendant will receive after found guilty.5 In the US judges, may be
elected or appointed and the judiciary is separated from the executive or legislative
branch of governments.
In Germany, lay judges often sit with professional judges in criminal trials
(Aronowitz, 2000). By way of contrast, in Japan, the jury system has by and large been
suspended and judges chosen on the basis of national level examinations adjudicate in
criminal trials (Moriyama, 2000).
As noted by Judge Richard Posner, impartial adjudication is central to well
functioning judicial system. As Posner put it: “the rules of the judicial process have been
designed both to prevent the judge from receiving a monetary payoff from deciding a
5 When criminal sentences were largely indeterminate (e.g. 5 years to life), judges and parole boards had substantial discretion. The
move to determinate sentencing (e.g., sentencing guidelines) decreased the amount of sentencing discretion available to judges.
case in a particular way and to minimize the influence of politically effective interest
groups on his decisions Posner, 1992, p.534). Posner’s first requirement suggests that for-
profit provision of adjudication is a non-starter. His second requirement suggests that
some insulation from electoral politics may be desirable.
But why do so many adjudication system involve, the lay public? One possible
justification for lay involvement in adjudication may be to obtain both public support for
and tempering of legal outcomes in the criminal arena. For example, it is well know that
juries will often not convict when the potential penalty is not in accord with community
feelings regarding what is appropriate.
Punishments for criminal offense range widely: death (capital punishment);
depravation of liberty (up to imprisonment for life); corporal punishment; control in
freedom (e.g., probation and parole); fines; warnings or admonitions; and community
service orders. According to the fifth United Nations Criminal Justice Survey (UNCJS),
which obtained information on crime and criminal justice for 1993/1994 for a wide
variety of countries, deprivation of liberty (i.e., imprisonment) was the most common
form of criminal punishment with 35% of cases resulting in this type of sentence. Fines
were the next most common sentence and were used in 33% of the cases (Newman, 1999,
The relative use of fines and imprisonment varies widely across countries. For
example according to the UNCJS survey, adjudicated criminal cases in Columbia always
result in imprisonment, 96% of criminal cases resulted in imprisonment in Greece, 92%
in Mexico and 60% in Italy. By way of contract, in Japan, 95% of adjudicated cases
resulted in fine, 82% in Myanmar, 79% in England and Wales and 70% in Egypt and
After carrying out a quantitative examination of sentencing practices, Shinkai and
Zvekic conclude that the level of development of the country, economic situation, or
region could not explain variations in sentencing practices. They conclude that cross-
national variations in sentencing patterns are best explained by the “availability and
acceptability of the sentencing options” (Newman, 1999, p. 91).
Economists have been interested in optimal sentencing since the work of Becker
(1968) and Stigler (1970). Economists generally conclude that fines should be preferred
whenever they can be imposed. See for example, Posner (1992, p. 227). The argument for
fines rests on their production of revenue for the state, victim or both and the high costs
of imprisonment. The major economic arguments for the use of imprisonment for the
core crimes rest on the fact that many who commit crimes are too poor to pay a fine that
would provide optimal deterrence. For a discussion see Kaplow and Shavell (1999). This
is so both because of their penury and because many of the core crimes have low
probabilities of the offender being penalized and, hence, would require very high fines if
optimal deterrence were to be achieved. For example, in New South Wales (Sydney,
Australia and surrounding communities), crime statistics for 1996 indicate that as a whole
those who break and enter buildings, steal cars, rob and assault others have only a 4%
chance of being convicted and less than a 1% chance of going to prison. The reason for
the low probability of apprehension and punishment for these crimes is mainly due to
victims’ failures to report the offenses to the police (54% of the offenses are reported),
failures of the police to record reported offenses (40% of crimes are recorded by the
police) and failures of the police to find the perpetrator (7% of the crimes are cleared by
the police) (Newman, 1999, p.75).
Another economic argument for imprisonment rather than fines is that
imprisonment quite successfully incapacitates offenders and fines do not. The social
benefit of this incapacitation depends upon the extent to which offenders will continue to
offend and on the elasticity of supply of offenders. This suggests that imprisonment of
consistently violent offenders will have higher social benefits than will imprisonment of
those who commit a “crime of passion.” Certainly it argues against the tendency in the
USA to use scare prison resources for drug dealers. As my son, who lives on the lower
east side of New York, says: “Mom, they are like cockroaches—as soon as one leaves
there is another to replace him.” Society achieves little decrease in drug dealing by
incarcerating drug dealers.
The arguments presented to this point suggest that wealthy offenders with little
likelihood of offending again should be fined not imprisoned. While this is largely the
case, it is not always so. Why? Before turning to possible economic arguments for
imprisoning such offenders, we will consider the reasons judges provide.
In United States v. Bergman (1976), a 64-year-old rabbi with an excellent
reputation for community service was convicted of fraudulently charging the government
for services rendered by nursing homes that he owned. In sentencing Bergman to a short
term of imprisonment, Judge Marvin Frankel carefully enunciated his reasoning. He
concluded that both general deterrence and equal justice required a prison sentence in this
It is also interesting to consider Judge Kimba Wood’s reasoning when sentencing
the US junk bond inventor Michael Milken to prison. She found that a prison term was
necessary in the Milken case to achieve general deterrence (i.e., the need to prevent
others from violating the law). Her reasoning is interesting. She found that prison
sentences are viewed as one of the most powerful deterrents to the financial community.
She also reasoned that crimes, like security fraud, which are hard to detect require greater
punishment in order to deter others from committing them (Kaplan, et. al., 1991, pp. 571-
In some ways, the arguments of Judge Frankel and Judge Wood are like those of
economists, but in other ways they are quite different. Both Bergman and Milken were
capable of paying large fines (indeed, Milken paid very large fines). Being barred for life
from working in the security industry effectively incapacitated Milken and Judge Frankel
found it unlikely that Bergman would ever offend again. This leaves only the economic
argument of optimal deterrence. Would it have been possible to achieve optimal general
deterrence by only fining Bergman and Milken. The Judges ruled not, but I suspect that
many economists would argue that given the wealth of these two criminals a large
enough fine would have effectively deterred others.
This leaves Judge Frankel’s second argument, equal justice. Economists are not
accustomed to thinking about equal justice, but are accustomed to thinking about
horizontal equity. As we have argued earlier, in the criminal justice setting, horizontal
equity might be seen as treating equally guilty parties equally. Sending the judgment-
proof poor to prison and allowing the equally guilty rich to pay a fine might strike some
economists as horizontally inequitable.
IV. PUBLIC AND PRIVATE PROVISION OF PREVENTION AND CRIMINAL
The previous section gave reasons why we might have public expenditures on
crime prevention and criminal justice. In this section we will focus on  how much
overall spending there should be on crime prevention and criminal justice  public
expenditures on crime prevention and criminal justice  individual crime prevention
activities and  the partial privatization of criminal justice activities.
Optimal level of expenditures on crime prevention and criminal justice
In many industrialized countries, an increasing amount of public and private
resources is devoted to crime prevention. What counts as prevention? Examples include,
specific crime prevention programs (e.g., juvenile delinquency, school, ex-offender job
training, rehabilitation and counseling programs), employment of security guards,
installation of locks, burglar alarms, CCTV systems and many other innovations in crime
preventive technology and policing. The socially or individually optimal level of crime
prevention is where the marginal benefit of reduction in crime equals the marginal cost of
extra prevention. However, as Freeman (1999) has noted, to estimate the marginal dollar
value of the reduction in crime due to any crime prevention policy is hard because of the
difficulties associated with measuring reductions in monetary and non-monetary costs
(e.g. reduced non-monetary loss from being victimized).
To calculate the tradeoff of the marginal value of the reduction in crime due to the
criminal justice system is even trickier than for prevention. The criminal justice system is
diverse and multifaceted. There are a host of agencies involved and at the margins these
agencies engage in much work having little to do with criminal justice. Thus, when the
youth services of an English town asked the simple question “did our interventions
against offending come to a profit or loss last year” the answer required a considerable
research effort. A criminal justice ‘audit’ was set in train to estimate costs of operating
the criminal justice system. For a discussion of such an audit see Shapland (2000).
Shapland notes that such work helps us to understand the respects in which criminal
justice is indeed a ‘system’. Efforts to gauge precisely the unit cost of each stage of
criminal justice (e.g., average costs to provide support to victims during the reporting and
investigation stage) have a wider significance. Only by comparing the use of resources
with such data can one see the effective priorities of the system, that is, upon what it
spends its money. Informed comparison of, say, spending on victim/offender mediation
compared with spending on refuges for rape victims, can then be made.
Interest in rigorous evaluations of crime prevention programs has increased in
recent years. Sherman et al., (1997), for example, introduces a scientific methods scale to
assess the methodological quality of evaluation studies in the US, and Goldblatt and
Lewis (1998) report similar research from the UK. Partially in response to this trend, the
UK Government established a three-year Crime Reduction Programme (1999-2002)
which included an assessment of the effectiveness and cost-effectiveness as a foundation
for setting priorities and allocating resources. Cost-effectiveness and cost-benefit
analyses of criminal justice agencies and programs require both estimates of the costs of
crime and the cost of the agency or program. As noted earlier, estimating the cost of
crime is difficult. However, estimates are becoming more common. For example, Brand
and Price (2000) provide estimates of the cost of crime for the UK, which include
monetary and non-monetary costs to victims. Cost-benefit analyses of crime prevention
programs have also been carried out in other industrialised countries (see references cited
in Brand and Price).
There appears to be a broad consensus in the international community on the
process required to ensure cost-effectiveness in reducing crime: increased collaboration
between institutions (see Walker and Sansfacon 2000). These institutions include
communities, families, schools, businesses, and government agencies, such as, law
enforcement, education, health, labour, social services, housing and urban planning. For
France, Walker and Sansfacon provide confirmation of the importance of coordination of
crime prevention programs between government agencies in reducing crime. The authors
describe “Local Prevention Contracts” in which mayors, chief prosecutors, police chiefs
and the national official for education sign contracts to support local crime prevention
projects covering most urban areas in France. Also, a community might end up with a
group of young workers who share the experience of work with police as safety and
security assistants or as social mediation agents.
In addition to increases in partnerships with other organizations, there has been
some interesting work on police tactics. On the one hand, the Kansas City preventive
patrol experiment (Kelling 1977) concluded that reduced police response time does not
reduce crime. On the other hand, community policing with a clear focus (e.g., directed
police patrol in crime hot spots) has shown substantial evidence of crime reduction in the
US (Sherman et al., 1997). Despite the fact that there is evidence that many different
crime prevention programs can effectively pevent crime (see Section V below), police
and prisons remain the two most fiscally important areas of criminal justice expenditure
in almost all developed countries.
In analyzing expenditures on criminal justice, it should be borne in mind that in
most countries, police, courts, and prisons are administered by many different agencies or
departments of government. Consequently, within countries it is difficult to identify
expenditures since financing is often conducted in separate government departments
which may be unrelated to criminal justice itself (see Newman, 1999, pp.137-138). It
should also be noted that while information problems concerning annual expenditures on
criminal justice exist within countries over time, comparisons at cross-national level are
sometimes even more problematic. These difficulties arise, in the main, from the way
different countries define crime, justice, and other relevant concepts (see Howard et. al,
2000, for a survey of comparative criminology issues)6.
According to the Bureau of Justice Statistics, the total amount spent by all levels
of government in the US in 1996 was $120 billion. Lindgren and Gifford (2000) report
that $53 billion was for police protection, $41 billion for corrections (e.g., prisons and
jails) and $26 billion for judicial and legal costs. In the U.S., government spending on
crime as a percentage of GDP was roughly 12
1% in 1996. Tables 1 and 2 summarize the
estimates of criminal justice expenditures for a number of major industrialized countries.
6 Clearly, differences in accounting practices may seriously affect capital and labour expenditure estimates reported by countries to
agencies, such as, UNCJS.
These cross-country aggregates are from van Dijk and de Waard (2000) and relate mainly
to 1997 and 1998.
The US and England and Wales have the highest levels of spending on crime as a
proportion of GDP with Denmark and France recording the lowest expenditure rates. A
striking feature of Table 1 is the relatively low level of police expenditure for the US
compared with the high level in England and Wales. This difference may reflect the high
ratio of private to public police in the US relative to England and Wales, but may also
reflect differential salary levels within and across these countries.
It is no surprise that the expenditure on prisons is much higher in the US than in
other countries given the substantial increases in the US incarceration rate over the last
three decades. For purposes of comparison, expenditures rates on prisons in France are
the lowest in the sample.
Estimate of Expenditures per mille of GDP, 1998 prices
Judiciary Prosecution Police Prison Total
7.37 1.76 10.53
8.79 1.15 12.60
Canada 1.20 0.34 7.23 2.48 11.25
Denmark 1.22 0.21 4.86 1.35 7.64
England and Wales 1.24 0.46 10.82 2.55 15.07
6.10 0.85 8.26
5.86 1.06 10.43
Netherlands 1.12 0.56 7.30 2.59 11.57
Sweden 1.67 0.42 6.07 2.18 10.34
US 2.91 0.47 6.75 5.51 15.64
Note: Estimates are italicized.
Source: van Dijk and de Waard (2000) p.49.
From Table 2 we note that the US is the biggest per capita spender with expenditures of
€436 per head, followed by Austria and England and Wales. The lowest per capita
spenders were France and Denmark with €180 and €184, respectively.
Estimate of Expenditures per capita in € , 1998 prices
Judiciary Prosecution Police Prison Total
160 38 229
203 26 290
Canada 28 8 169 58 263
Denmark 30 5 117 32 184
England and Wales 23 9 205 49 286
132 19 180
137 25 245
Netherlands 23 11 151 54 239
Sweden 33 8 119 43 203
US 81 13 188 154 436
Note: Estimates are italicized. Fixed exchange rates for € zone countries and PPP 1998 for other countries.
Source: van Dijk and de Waard (2000) p.50.
During the past two decades, the number of police per head of population for both
industrialized and developing countries has increased, although the greatest increases
have been seen in industrialized countries. As Newman (1999) points out, there exists a
strong positive correlation between expenditure on criminal justice and economic wealth
(as measured by GDP per capita). Table 3 focuses on police expenditure per capita and
GDP per capita. In general, we see that richer countries spend more on policing.
However, Japan is one clear exception with a very high GDP but low police expenditures.
Police Expenditure and GDP, 1994 (per capita US $)
Country GDP (US $)
Police (US $)
Police as a
percentage of GDP
Colombia 1847 18.72 1.01
Costa Rica 2463 7.42 0.30
Croatia 3867 20.57 0.53
Cyprus 9754 136.59 1.40
Denmark 28245 145.28 0.51
Finland 19048 112.23 0.59
France 24608 148.90 0.61
Greece 7465 60.01 0.80
Hong Kong 22590 185.65 0.82
Hungary 4072 6.09 0.15
India 309 0.20 0.06
Japan 36782 18.40 0.50
Jordan 1095 15.42 1.41
Madagascar 208 0.05 0.02
Malta 7394 77.09 1.04
Netherlands 21536 204.09 0.94
Romania 1274 4.13 0.32
Saint V. & Grenadines 2248 41.45 1.84
Singapore 23556 100.94 0.43
Slovenia 7206 98.13 1.36
Spain 12201 27.73 0.23
Sweden 22499 157.29 0.70
Switzerland 36096 299.53 0.83
Turkey 2227 9.66 0.43
Source: Adapted from Newman (1999) p.302
The table above tells us that richer countries spend more on police per head. However,
rough calculations suggest that developing countries (e.g., Colombia, Cyprus, Jordan,
Saint V & Grenadines and Solvenia) spend more on police as a % of GDP relative to
Rapidly growing prison populations in many countries has led to an upsurge of
interest in discerning the impact of this costly increase on crime rates. For example,
recent work in the US, using either state-level panel data on crime rates from the FBI’s
Uniform Crime Reports (e.g., Levitt, 1996) or time-series data (e.g., Witt and Witte,
2000), finds increased imprisonment to be associated with significant declines in the
reported crime rate. It is important to note that this work considers the effect of increased
imprisonment on crimes reported to the police. As noted earlier, reported crime can
change even when actual crime does not. For example, victims can decide to report more
or less crime to the police and the police can decide to record more or less of the crime
that they uncover.
As of midyear 1999, the US had incarcerated 1,860,520 individuals in its prisons
and jails. This represents an incarceration rate of 1 in every 147 US resident. Estimates of
the annual cost of locking up an inmate in the US can be found in Donohue and
Siegelman (1998). For example, Donohue and Siegelman (p. 5) estimate that the annual
cost of incarcerating an additional inmate is approximately $36,000 (in 1993 dollars).
Although this estimate includes cost of building, occupying a prison cell and lost
legitimate wages, it ignores a number of social benefits (e.g., the benefit from seeing an
individual punished) and social costs (e.g., effects of imprisonment on future legitimate
Not surprisingly, there are vast differences in expenditure on prisons between
developed and developing countries. As well as differences in expenditure on police and
courts between developed and developing countries, richer countries also tend to spend
7 In some countries, the military assumes some police functions. For example, in the US, the military was used to help fight the “war
more on prisons, although Japan again, with its high GDP spends relatively little on
prisons (Newman 1999). Table 4 shows UNCJS figures for annual public expenditures
per convicted prisoner. These data are derived by multiplying the expenditure (salaries
and fixed assets) on corrections (penal and correctional institutions) reported by each
country in local currency by an exchange rate and then dividing by the number of
convicted adult prisoners reported in 1994. With the exception of Northern Ireland, which
has its own unique characteristics, Switzerland is currently one of the highest spenders,
alongside Sweden, US, Denmark and England and Wales. Interestingly, these are the
same countries identified in Figure 1, p. 4, as having stronger rules of law.
Annual Expenditure per Convicted Prisoner, 1994 (US $)
Northern Ireland 158197
England and Wales 61721
Hong Kong 28341
Rep of Korea 10122
Czech Rep 8903
*Brunei Darussalam 4253
Costa Rica 1923
Source: Newman (1999) p.142
An examination of expenditures per convicted prisoner indicates some interesting facts.
Japan has substantially lower levels of spending on prisons per head of the population,
while at the same time, spends a high amount per prisoner. Another fact that stands out is
that developing countries have lower levels of prison expenditure per prisoner compared
with industrial countries. These particular comparisons are obviously sensitive to the
precise choice of day of year, given that the number of admissions to prison is not taken
into account (see Newman, 1999, p.337).
Individual Efforts to Prevent Crime
In general, an individual will purchase crime prevention goods and services when
the cost of prevention is less than the expected benefits from prevention. One aspect of
the debate over the modes of crime prevention is the separation between private and
public expenditures. There are a number of explanations as to why certain individuals
may only be concerned about private expenditures. The failure of government programs
to stem the growth in crime may provide one example, but an alternative explanation may
lie in terms of how individual objectives are determined. The median-voter model,
originally developed in the political sciences, may be applied to shed light on the level of
private prevention expenditures. In democracies, things like police services will be set at
the level desired by the median voter (generally considered to be the voter with median
income). Individuals and firms with above median income may quite rationally choose to
increase their level of protection by buying in the private market. Under such
circumstances, it could be argued that richer individuals and firms purchase relatively
more protection, because they personally stand to gain more from this than from the
alternative low-protection strategy.
Expenditures by individuals and private organizations on crime prevention are
more difficult to estimate than public expenditures. In 1992, the latest year in the US for
which we have a benchmark input-output table, private household purchases of detective
and protective services was $944 million and purchases of security systems services was
$1301 million. Clearly, these are not the only private purchases related to crime
prevention and criminal justice. Purchases of legal services by private households was
about $44 billion in 1992, but we have no information on what part of this was related to
criminal cases. There are also expenditures on modifications to existing structures (e.g.,
bars on windows) car alarms, and other anti-theft devices that do not show up explicitly
in the estimates. The above private expenditures do not include expenditures by
businesses and other organizations. Such organizations have substantial crime prevention
and apprehensions expenditures as is clear from a trip to any major company or
university. Unfortunately, such estimates are not readily available in the UK due to the
absence of systematic accounting of private crime prevention expenditures.
Cooter and Ulen (2000) on p. 458, quote figures for US private expenditures on
crime prevention in 1993 of $65 billion. Sources cited in Anderson (1999) estimate that
expenditures on private protection in 1993 are $69 billion. Philipson and Posner (1996)
cite a yet higher estimate of $300 billion (includes expenditures by businesses and other
enterprises on security guards and other measures of self-protection). Laband and
Sophocleus (1992) provide a similar estimate.
Anderson concludes that the aggregate burden of crime, taking into account the
value of lost property, transfers, and losses to victims of crime (e.g., worth of assets from
victims, lost productivity, medical expenses, and diminished quality of life), is $1,705
billion. This estimate seems excessive given it is approximately one-fifth of the US GDP
reported in 1999.8
It would appear that private and public anti-crime initiatives can be either
substitutes or complements. Philipson and Posner, for example, show that the proportion
of homes with burglar alarms in a state falls with improved public sector crime protection
schemes. Ayres and Levitt (1998) find that the introduction of a Lojack system for
recovering stolen cars reduces overall car crime.
Measures of the cost of crime and of private expenditures on crime prevention
and criminal justice are sparse and, yet, such numbers are central to being able to talk
intelligently about either the public/private tradeoff or the optimal level of overall
expenditures on crime prevention and criminal justice. This is an area ripe for detailed
and carefully done empirical work.
Partial Privatization of Criminal Justice Activities
What role does the private sector play in the criminal justice system? The
American example is perhaps the most dramatic in the variety of private sector activities
that characterize its criminal justice system. Benson (1998) provides a comprehensive
account of the public sector contracting out to the private sector. Examples of this partial
privatization include police services, drug treatment facilities, airport security, prisons
and correctional facilities.
It is now commonplace for private firms to provide a whole range of services
previously supplied by governments. Corrections Corporation of America and
Wackenhut Corrections Corporation, for example, provide correctional, security and
other related services to government agencies around the world. Wackenhut contracts
include, security at the US Embassy in El Salvador, supplying the entire police force for a
nuclear power plant in Illinois and providing correctional facilities in New Zealand.
Benson provides evidence to show that private security and community policing
initiatives have been extremely successful in reducing crime. Examples include, private
residential streets, patrols and neighborhood watch, the deterrent effect of gun
8 We have been unable to find estimates of private expenditure on crime prevention goods and services for other countries.
ownership9, and technology (Lojack example mentioned above). In addition to these
private sector efforts, he argues that the criminal justice system should employ more
resources in giving reparation for loss or injury inflicted to the victims of crime. Benson
notes that private sanctions imposed by firms (e.g. firing an employee who steals from
the firm) are now being substituted for public sector criminal prosecution.
The contracting out by governments around the world of prison management
services to private companies has grown rapidly over the last decade. As figure 6
documents, the total number of prisoners held in private facilities rose sharply from
15,300 in 1990 to 145,160 in 1999. For example, the number of private prisoners in the
US has grown from about 1200 in 1985 to 122,871 at the end of 1999, which represents
6.6% of total prison population. Estimates of the number of private prisoners in other
9 The deterrent effect of private gun ownership is very controversial. See Cook & Ludwig (2000). As a whole Benson’s book has been
the subject of considerable controversy. It would be very useful to have scholars with different perspectives consider the benefits and
costs of privatisation of crime prevention and criminal justice.
countries, at end of 1999, are Australia (7459), England and Wales (7161), Netherlands
(737), New Zealand (384), Scotland (500) and South Africa (6048). With the exception
of Australia, these totals are very small relative to the total number of prisoners.
The US, Australia and the UK have been the main countries to experiment with
private prisons. At the end of 1999, the number of privately managed secure adult
facilities in the US, Australia and the UK were, 158, 15 and 10, respectively. Corrections
Corporation of America and Wackenhut Corrections Corporation have the largest global
market share of contracts to run private prisons.
Owing perhaps to the rise of privatizing governmental functions and to the claim
that private prisons are cheaper per prisoner than public prisons, the issue of privatization
of prisons has become highly contentious (see, for example, Shaw 1994, Biles 1997, Hart
et al. 1997, and Benson 1998). Arguments for governments to manage prisons generally
rest on the need to maintain a high level of standards in the quality of prison services and
in the behavior of prison employees. Contracts for private prison operators would tend to
Rated Growth Capacity of Private Secure Adult Correctional
1990 19911992 19931994 199519961997 1998 1999
Source: Thomas (2000) http://web.crim.ufl.edu/pcp/census/1999/Figure1.html
be quite incomplete; privately contracted firms would have strong incentives to lower
standards in order to minimize costs (Hart, et al, 1997). There are a number of recent
examples where this has actually happened. A prison in Louisiana was recently taken
away from the private sector because of unacceptable conditions. Others argue that it is
not necessarily the contractual incompleteness that has an adverse effect on quality, but
rather the inability or unwillingness of the client to enforce contract terms (Domberger
and Jenson, 1998).
There has been a substantial growth in private policing in many countries. As
Newman observes, “While private policing has a long history in industrial countries, it is
also becoming a major growth industry in emerging market economies” (1999, p.126).
Sources cited in Newman (1999) estimate that the number of security guards in
Singapore (with a population of 2.6 million) is approximately 15,000 to 20,000, which is
at least twice the police strength. A high ratio of private to public police is also found in
other developed economies. For example, the US has a security guard (or officer)/police
ratio of three to one (see Newman p.127).
V. WHAT WE KNOW ABOUT THE RELATIVE EFFECTIVENESS OF
CRIME PREVENTION AND CRIMINAL JUSTICE
Just how effective are specific crime prevention programs? What are the major
benefits of incarceration? In this section we do two things. First, we survey briefly a
range of crime prevention strategies that seem to work in the US. Second, we look at the
arguments that have been put forward relating to the economic issues surrounding the
benefits of incarceration. Standard texts on law and economics (e.g., Cooter and Ulen,
2000) discuss four types of social benefits derived from imprisonment, namely,
retribution, rehabilitation, deterrence and incapacitation. Given the difficulty of
measuring the first two, we shall concentrate on deterrence and incapacitation. It should
be noted, however, that much of the empirical literature does not distinguish between
deterrence and incapacitation effects. In the case of imprisonment, separating
incapacitation from general deterrent effects is difficult since the two are jointly
Specific Crime Prevention Programmes
Potential effective crime prevention programmes could take a number of forms.
Sherman et al., (1997), writing from a US perspective, points to a number of successful
programmes. These include, community based mentoring and afterschool recreation
programmes, intensive work with at-risk families with young children, intensive
residential training programmes for at-risk youth, extra police patrols in high-crime hot
spots and a number of situational crime prevention schemes, such as neighbourhood
watch and building and community redesign programmes.
The recently completed evaluation of Children at Risk (CAR) programme found
that youths in the treatment group had participated in more social and educational
activities, exhibited less antisocial behaviour, committed fewer violent crimes and used
and sold fewer drugs than did youths in the control group (see Harrell, Cavanagh and
Sridharar 1999). CAR was a drug and deliquency prevention program for high risk
adolescents between the ages of 11 and 13 who lived in five cities (Austin, Bridgeport,
Memphis, Savannah, and Seattle). The programme consisted of eight components
considered key to comprehensive delinquency prevention: case management, family
services, educational services, after-school and summer activities, mentoring, incentives,
community policing and enhanced enforcement, and criminal-juvenile justice
intervention. One interesting result was that the positive effects of the programme on
drug use, crime, and risk factors were not generally observed at the end of the
programme. This may indicate that CAR was simply a secondary prevention programme
when youths got into trouble.
The Job Corps programme in the US has long been a central part of the federal
government efforts to provide training for disadvantaged youths. Recent evidence
suggests that participation in the programme significantly reduced arrest and conviction
rates, as well as time spent in jail (see Schochet, Burghardt and Glazerman 2000). The
arrest rate was reduced by about 6 percentage points and the impacts on arrest rates were
very similar across male and female subgroups.
The evaluation of the effectiveness of private crime prevention activities is much
more difficult to find. However, given the recent epidemic of households and firms
buying locks or alarms or other forms of protection in response to the threat of crime,
researchers have started considering how these activities can have negative or positive
spillover effects for the neighbourhood community. For example, Ayres and Levitt
(1996) conclude that the Lojack system for recovering stolen cars (hidden radio-
transmitter that enables the police to locate the stolen vehicle) provides an example of a
positive externality due to a general deterrent effect.
Related to the above initiatives are those of how communities organise
themselves when allocating resources to crime prevention programmes. Hawkins (1999),
for example, explores the implications of a “Communities That Care” system where
prevention science is used to guide the type of prevention policies suitable according to
profiles of risk and protection.
There is a reasonably large theoretical and empirical literature in economics that
considers both the deterrent (specific and general) and incapacitative effect of
imprisonment. For example, economists and others have sought to discern if increased
imprisonment lowers the crime rate? How does the deterrent effect of formal sanctions
Following Becker’s theoretical work, much empirical work by economists has
focused on the role of the criminal justice system in determining criminal activity. In
short, economists have argued, using both theory and empirical work, that if the cost of
crime is raised, by increasing the probability of apprehension or imposing more or longer
sentences, less of it will occur.10 Deterrence refers to the effect of possible punishment on
individuals contemplating criminal acts11. Deterrence may flow from both criminal
justice system actions and from social actions (i.e., the negative response of friends and
associates to criminal behavior). To date, attempts by economists to measure deterrent
effects have concentrated on the effects of the criminal justice system while work by
10 Actually the deterrence hypothesis was widely explored both empirically and theoretically prior to economists re-entry to the study
of crime in the late 1960s.
11 Marginal deterrence refers to a situation where individuals commit less harmful rather than more harmful acts if expected sanctions
rise with harm (see Polinsky and Shavell, 1999).
sociologists have concentrated on “social sanctions.” See Nagin (1998) for a survey of
The potential criminal’s perceptions regarding social sanctions are difficult to
measure, although work on peer group (e.g., Evans, et al, 1992) and community effects
(e.g., Sampson and Groves, 1989) may be able to capture some aspects of perceived
social sanctions. In terms of the effects of education, a consistent finding is that students
who attend schools with strong ethical values (e.g., parochial schools) offend less than
students attending modern urban high schools (see, e.g., Tauchen et al. 1994). This effect
may stem from higher levels of social sanctions against crime in schools with strong
ethical standards or from the better family and community settings of at least some
In an interesting paper, Williams and Sickles (1999) provide an extension of
Ehrlich (1973) by including an individual’s social capital stock into his utility and
earnings functions. Social capital, including things like reputation and social networks, is
used as a proxy to account for the effect of social norms on an individual’s decision to
participate in crime. This assumes that the stigmatism associated with arrest depreciates
an individual’s social capital stock. Williams and Sickles clarify this point further by
arguing that employment and marriage create a form of state dependence, which reduces
the likelihood of criminal involvement. In other words, an individual with a family, job or
good reputation has more to loose if caught committing crimes than those without such
attachments. Dynamics arise from current decisions affecting future outcomes through
the social capital stock accumulation process. Their main result is that criminals behave
rationally in the sense that they account for future consequences of current period
Glaeser, Sacerdote and Scheinkman (1996 p. 543) make a similar point in the
context of family structures, claiming “the average social interactions among criminals
are higher when there are not intact family units. The presence of strong families
interferes with the transmission of criminal choices across individuals”. The importance
of family structure (along with other variables such as deterrence and returns to crime) in
explaining urban crime is also the key point of Glaeser and Sacerdote (1999).
12 Many of these insights into the dynamics of crime were originally discussed in Williams PhD dissertation (1995)
In the literature deterrence is broken into two components. The first component,
called specific deterrence, encompasses the effect of punishment on the individual
punished. The second component, called general deterrence, encompasses the effect of
punishment on the general public. Specific deterrence is generally reflected by including
measures that reflect the individual’s past experience with the criminal justice system
(e.g., Witte 1980; or Trumbull 1989). The implicit assumption is that offenders from their
perceptions regarding possible punishment based on their own experience with the
criminal justice system13. For example, the offender’s perceived probability of arrest
might be proxied by the ratio of his past self-reported offenses to arrests and his
perceived punishment as some sort of average of the punishments he has received in the
past. There is an important potential difficulty in using this type of specific deterrence
measure. If there is autocorrelation in criminal behaviour, these measures of specific
deterrence will be correlated with the error term in the crime equation. One might
instrument these variables by using community-level or peer group measures.
It has proven much more difficult to obtain reasonable measures of general
deterrent effects. As an example, consider the probability of arrest. In a standard model of
criminal choice, an individual’s probability of arrest depends upon his level of criminal
activity, his ability to avoid arrest, and exogenous factors related to the criminal justice
system. When contemplating a crime, the individual is faced with a schedule of
probabilities that relates the nature and extent of his criminal activity to the probability of
arrest. See Cook (1979) or Tauchen et al. (1994) for a discussion. An analogy would be
to a taxpayer who, when making her labour supply and tax reporting decisions, is faced
with a schedule that relates reported income to the schedule of tax rates.
Just as there is no single tax rate, there is no single probability of arrest. There is a
different probability of arrest for each and every possible set of criminal choices. For
example, we would expect that for a given individual the probability of arrest would be
much higher for robbery than for petty theft.
13 We know little about how individuals form their perceptions of likely sanctions if they offend, although Paternoster and his
colleagues (see, e.g., Nagin and Paternoster, 1991) have done interesting empirical work, and Sah (1991) has developed a model for
the perceived probability of punishment. In a study of institutionalised young adults (college students and prison inmates), Lattimore
et al. (1992) find that individuals transform probabilities when making risky choices. Risk seeking is common over long-shot odds,
and subjects are less sensitive to changes in midrange probabilities than is assumed by expected utility models.
Changes in criminal justice policy or in the level of criminal justice resources
alter the probability schedule facing a potential criminal. For example, an increase in
criminal justice resources such as that contained in the 1994 US Crime Bill might raise
the probability of being arrested for each criminal act, that is. It might cause the schedule
relating the probability of arrest to criminal activity to shift up. The “war on drugs”
caused certain sections of the probability schedule (the sections associated with drug
offenses) to shift up and other sections to shift down (the sections associated with
violence offenses). It is these types of exogenous changes in the criminal justice system
that should be used to reflect deterrent effects and not a community-level probability of
arrest. This approach to representing deterrence has been used by Block et al. (1981) and
Tauchen et al. (1994).
There are a number of practical problems that arise in testing for deterrent effects.
In particular, we consider three estimation issues: measurement error, endogeneity and
Models of criminal behaviour are usually estimated using official reported crime
statistics. Such recorded offences are influenced both by victims' willingness to report
crime and by police recording practices and procedures. At the level of the individual
police department, both administrative and political changes can lead to abnormalities in
reported data or to failures to report any data. For example, the measurement error in
crime rates may arise because hiring more police leads to more crimes reported.
Consequently, estimates derived from regressing crime rates on the number of police (or
on arrest rates) may be severely distorted by the impact of measurement error. Until quite
recently, measurement error was not widely considered in the economics literature.
However, the importance of the issue and potential solutions were considered very early
by Carr-Hill and Carr-Hill (1972) and Carr-Hill and Stern (1973).
More recently, Corman and Mocan (2000) report that complaints to the police for
murder, robbery, burglary and motor vehicle theft “decline in response to increases in
arrests.” They use detailed information from the New York Police Department and
modern time series estimation techniques. By using data for a single police department,
they avoid some of the measurement error inherent in studies that use cross department
statistics. However, changes in victim reporting behaviour are not considered.
Murder is the one crime for which police reports are quite good. Corman and
Mocan report that the elasticity of murders reported to the police with respect to the arrest
rate is approximately -.3. A number of authors have noted recently that much of the
decline in murder rates in developed countries has stemmed from a decrease in domestic
assaults. Corman and Mocan do not include measures to reflect the change in public and
police attitudes regarding domestic violence in their equations. In general, economists’
work on crime are only beginning to incorporate family and community effects.
Somewhat unsurprisingly, economists have concentrated their attention on the
possibility that crime and sanctions are jointly determined. The main point is that
increases in sanctions may cause decreases in crime, but increases in sanctions may also
be a response to higher crime rates. Since the 1970s there has been a considerable effort
to find instruments (i.e. exogenous factors) to identify the effects of sanctions on the
supply of crime. For example, Levitt (1996) uses instrumental variables to estimate the
effect of prison population on crime rates. Prison-overcrowding litigation in a state is
used as an instrument for changes in the prison population.
In order to identify the effect of police on crime, Marvell and Moody (1996) and
Levitt (1997) proposed different procedures. Marvell and Moody are concerned with the
timing sequence between hiring police and crime. Using lags between police levels and
crime rates to avoid simultaneity, they test for causality in the spirit of Granger (1969).
Although they find Granger causation in both directions, the impact of police on crime is
much stronger than the impact of crime on police. In a recent paper Levitt (1997) uses the
timing of elections (when cities hire more police) as an instrumental variable to identify a
causal effect of police on crime. He finds that increases in police, instrumented by
elections, reduces violent crime, but have a smaller impact on property crime. Levitt does
not consider the impact of elections on either victim or police reporting behaviour.
A substantial problem that has been ignored in the vast majority of empirical
studies is nonstationarity of crime rates. A time-series is said to be nonstationary if (1) the
mean and/or variance does not remain constant over time or (2) covariance between
observations depends on the time at which they occur. In the US, the index crime rate
appears strongly nonstationary (see, for example, Witt and Witte 2000). Here, the authors
have attempted to estimate and test a model using time-series cointegration techniques.
The empirical results suggest a long-run equilibrium relationship between crime, prison
population, female labour supply and durable consumption.
against members of society outside prison14. Therefore, unless there is an infinitely elastic
Incapacitation refers to when an offender behind bars cannot commit new crimes
supply of criminals, incapacitation will reduce crime over what it would otherwise have
been. However, as argued in section II with respect to drug dealers, if the supply of
offenders is reasonably elastic the incapacitation effect may be very small (see Freeman
1996). In addition, for some individuals, incarceration may only affect the timing rather
than the total number of crimes they commit.
These uncertainties of the effects of prison on crime are illustrated in a recent
report by the Justice Policy Institute (2000) on incarceration and crime trends in Texas.
As of year end 1999, there were 706,600 Texans in prison, jail parole or probation, the
largest population of inmates under the jurisdiction of its prison system in the U.S.
However, there is little evidence that Texas’ severe correctional system is responsible for
the fall in crime; Texas crime rates have not experienced the declines witnessed in other
parts of the U.S., where prison growth has been much slower.
The basic intuition of the optimal sanction is that individuals should be put in
prison and kept there as long as the expected net harm exceeds the costs of imprisonment.
Polinsky and Shavell (1999) argue that prison should only be used to incapacitate
individuals whose net harm is relatively high. In addition, these authors highlight two
points about incapacitation. First, since expected harm caused by individuals usually
declines with their age, it may be worth having fewer older people in prison. Second, past
behavior is the best predictor of future behavior. Thus, the criminal justice system should
impose a prison sentence on someone who has committed a harmful act rather than
incapacitating someone who has the potential to commit a crime. In practice, you have to
commit a crime to go to prison.
While the theory of optimal incapacitation policy is well established (see Shavell
1987), the empirical work faces some difficulties. The central difficulty is how to
14 As discussed in Polinsky and Shavell (1999), incapacitation can take forms other than imprisonment e.g., loss of a driver’s license
prevents an individual from doing harm while driving.
separate the deterrent and incapacitation effects associated with imprisonment. One
approach, discussed at length in Ehrlich (1981), is to compare regression estimates of the
actual effect of imprisonment on crime with theoretical estimates of maximum
incapacitation effects. Some authors, such as Levitt (1998), seek to identify some
observable substitution effects of the probability of arrest for one crime on the incidence
of a substitute crime. Other authors, such as Kessler and Levitt (1999), have used
California’s adoption of “three strikes and you’re out” sentencing rules to separate the
deterrence and incapacitation effects of punishment.
Recently, some researchers have focused their attention on whether sanctions
depend on offense history. Glaeser and Sacerdote (2000) provide evidence to show that
repeat offenders are more likely to receive longer sentences. Offenders with a high
expected probability of recidivism are more likely to commit crimes in the future and
thus more likely to be worth incapacitating by imprisonment. Glaeser and Sacerdote use a
variety of data sources to examine the sentences given to murderers in the U.S. They find
that sentences are longer when there is a greater value to incapacitation or greater
deterrence elasticity. However, contrary to the predictions of the economic model of
optimal punishment, they find victim characteristics are important in explaining
sentencing among vehicular homicides (e.g., drivers who kill women get 56 percent
longer sentences, whereas, drivers who kill blacks get 53 percent shorter sentences).
VI. CONCLUSIONS AND SOME SUGGESTIONS FOR RESEARCH
Crime is pervasive. However, the extent of crime is hard to measure. Cross
nationally, we have the best measures for murder. However, even for murder estimates
from different sources can vary widely (see Figure 4). Given differences in criminal laws,
defendant protections and statistical systems, broad measures of law abidingness (e.g.,
those developed by the World Bank), and cross-national victimization and self-report
surveys offer the best hope of comparing the level of other types of offenses across
countries. Crime statistics coming from the criminal justice system (e.g., offenses
reported to the police) provide valuable information, but do not generally provide reliable
estimates of crime either across countries or across time.
To date, much work by economists on crime has used criminal justice statistics
and interpreted these statistics as measures of the underlying level of crime. This first
generation work has been valuable, but it is now time to move forward. Two directions
appear particularly promising. First, in order to interpret criminal justice data, we need to
know more about the behavior of both crime victims and the criminal justices system. For
example, in order to properly interpret results that use police reports as a measure of
crime, we need to understand both victim reporting behavior and police recording
practices. Second, work seeking to estimate the deterrent effects of the criminal justice
system could benefit from using sources of data on crime other than data from the
criminal justice system (e.g., victimization surveys). Studies that use a number of
indicators of crime (e.g., multi-indicator models) are likely to provide more meaningful
results than studies that use any single indicator.
To date studies of crime by economists have been focused on relatively few issues
(e.g., the deterrent effect of police resources and imprisonment). There have been
relatively few studies of the way in which such important criminal justice system entities
as prosecutors’ offices and public defenders operate.15
Most, if not all, criminal justice systems operate with budgets that are not
adequate to fully process all cases. To deal with this “overcrowding,” one or more
agencies are given discretion in how they handle cases. Different criminal justice systems
grant differing amounts of discretion to different agencies at different points in time. For
example, in the US during the 1960s, judges and parole boards had substantial discretion.
The determinate sentencing movement limited the discretion of judges and parole boards.
Discretion was not eliminated; it was simply shifted to prosecutors from judges and
parole boards. How much discretion in criminal justice system is optimal? Where should
the discretion be lodged?
Public finance economists have provided valuable work on optimal taxation. As
far as we are aware, there has been little work on optimal criminal law or optimal
criminal justice funding. What acts should we criminalize? What is the optimal level of
15 There has been some work by economists in the US on plea bargaining.
funding for the police, prosecutors, courts and corrections? What should the split between
private and public spending be? While private prisons have received some attention,
private policing has received less.
Much criminal justice research has focused on preventing crime through the
punitive actions of the criminal justice system (e.g., deterrence, incapacitation). Another
line of research has focused on preventing crime by working with high-risk youth. The
two lines of research are not well integrated. Yet, an optimal portfolio of crime
prevention strategies requires a combination of punitive and supportive efforts.
Anderson, D.A., (1999), ‘The Aggregate Burden of Crime’, Journal of Law and Economics, vol. XLII, pp.611-642.
Argersinger v. Hamilin 407 U.S. 25 (1972).
Aronowitz, A. (2000), ‘Germany’,
gov/bjs/pub/ascii/wfbcjger.txt, accessed July 7, 2000.
Ayres, I. and Levitt, S.D., (1998), ‘Measuring positive externalities from unobservable victim precaution: and empirical
analysis of Lojack’, Quarterly Journal of Economics, 113, pp.43-77.
Becker, G. (1968), ‘Crime and punishment: an economic approach’, Journal of Political Economy, vol. 76, pp.169-217.
Benson, B.L. (1998), ‘To Serve and Protect’, New York, New York University Press.
Biles, D. (1997), ‘Private prisons: welcome or not?’, Australian Journal of Forensic Sciences, vol. 29.
Block, M., Nold, F. and Sidak, J. (1981) “The Deterrent Effect of Antitrust Enforcement”, Journal of Political
Borricand, Jacques. (n.d.). France: World Factbook of Criminal Justice Systems. retrieved from the World Wide Web
on October 26, 2000: http://www.ojp.usdoj.gov/bjs/pub/ascii/wfbcjfra.txt
Brand, S. and Price, R. (2000) “The Economic and Social Costs of Crime”, Home Office Research Study No.217
Carr-Hill, G. and Carr-Hill R. (1972) “Reconviction as a Process”, British Journal of Criminology, 12:35-43.
Carr-Hill, R.A. and Stern, N.H. (1973). `An Econometric Model of the Supply and Control of Recorded Offences
in England and Wales', Journal of Public Economics, l.2, pp.289-318
Chong, Chee Kin. (2000, May 25). Cisco Escorts and Private Doctors for Convicts. The Straits Times (Singapore).
42 Retrieved November 5, 2000, from Lexis-Nexis database on the World Wide Web: http://web.lexis-nexis.com.
Cook, P.J. (1979), “The Clearance Rate as a Measure of Criminal Justice Effectiveness”, Journal of Public Economics,
Cook, P.J.and Ludwig, J. (2000), “Gun Violence: The Real Costs”, New York, Oxford University Press.
Cooter, R. and Ulen, T. (2000). Law and Economics, 3rd edition, Reading, MA: Addison-Wesley.
Corman, H. and Mocan, H.N. (2000), “A Time-Series Analysis of Crime and Drug use in New York City”, American
Domberger, S. and Jensen, P., (1998), ‘Contracting out by the public sector: theory, evidence, prospects’, Oxford
Review of Economic Policy, vol.13 pp. 67-78.
Donohue, J.J., and Siegelman, P. (1998), ‘Allocating Resources Among Prisons and Social Programs in the Battle
Against Crime’, Journal of Legal Studies, vol. XXVII, pp. 1-43.
Ehrlich, I. (1973) “Participation in Illegitimate Activities: A Theoretical and Empirical Investigation”, Journal of
Political Economy 81:521-565.
Ehrlich, I. (1981) “On the Usefulness of Controlling Individuals: An Economic Analysis of Rehabilitation,
Incapacitation and Deterrence”, American Economic Review, 71:307-22.
Evans, W., Oates. W. and Schwab, R. (1992), “Measuring Peer Group Effects: A Study of Teenage Behavior”, Journal
of Political Economy, 100, pp.966-91.
Freeman, R. (1996), “Why do so many young American men commit crimes and what might we do about it?”, Journal
of Economic Perspectives, 10:25-42.
Freeman, R. (1999), “The Economics of Crime”, Ch. 52 in Ashenfelter, O. and Card, D. (ed.), Handbook of Labor
Economics 3: North Holland.
Friedman, D. (2000), Law’s Order, Princeton, NJ: Princeton University Press.
Gideon v. Wainwright 372 U.S. 335 (1963).
Glaeser, E.L. and Sacerdote, B. (1999), “Why is There More Crime in Cities?”, Journal of Political Economy, 107,
Glaeser, E.L. and Sacerdote, B. (2000), “The Determinants of Punishment: Deterrence, Incapacitation and Vengeance”,
NBER Working 7676.
Glaeser, E.L., Sacerdote, B. and Scheinkman, J. (1996), “Crime and Social Interactions”, Quarterly Journal of
Economics, pp. 507-548.
Goldblatt, P. and Lewis, C. (ed) (1998) Reducing Offending: An Assessment of Research Evidence on Ways of Dealing
with Offending Behaviour, London: Home Office Research Study 187.
Granger, Clive 1969, Investigating Causal Relations by Econometric Models and Cross-Spectral Methods,
Groves, W., Newman G. and Corrado C. (1987), ‘Islam, modernization and crime: a test of the religious ecology
thesis,’ Journal of Criminal Justice, vol. 13, pp. 23-29.
of Criminal Justice Systems,
Harrell, A., Cavanagh, S. and Sridharar S. (1999), Evaluation of the Children at Risk Program: Results 1 Year After the
End of the Program, U.S. Department of Justice, Office of Justice Programs, Washington DC.
Hart, H. L. A. (1994), ‘The Concept of Law’, Oxford: Oxford University Press.
Hart, O., Shleifer, A. and Vishny, R. (1997), ‘The proper scope of government: theory and an application to prisons’,
Quarterly Journal of Economics, vol. 112, pp. 1163-1202.
Hawkins, J. D. (1999), ‘Preventing crime and violence through communities that care’, European Journal on Criminal
Policy and Research, vol. 7, pp. 443-458.
Howard, G.J., Newman, G. and Pridemore, W.A. (2000), ‘Theory, Method, and Data in Comparative Criminology’ in
Duffee, D. (eds), Criminal Justice 2000, Measurement and Analysis of Crime and Justice, Vol. 4, pp.139-211,
National Institute of Justice.
Justice Policy Institute (2000) “Texas Tough?: An Analysis of Incarceration and Crime Trends in the Lone Star State”
Jurgen_Tas, J., Terlouw, G. and Klein, M. (eds) (1994), Delinquent Behavior of Young People in the Western World,
Amsterdam: Kugler Publications.
Kaplan, J., Skolnick, J. and Feeley, M. (1991), Criminal Justice, 5th Edition, Westbury, NY: Foundation Press.
Kaplow, L. and Shavell, S. (1999), ‘Economic Analysis of Law,’ NBER Working paper 6960.
Kaufman, D., Kraay, A. and Zoido-Lobaton (1999a), ‘Aggregating Governance Indicators’, Working Paper, World
Kaufman, D., Kraay, A. and Zoido-Lobaton (1999b), ‘Governance Matters, Working Paper, World Bank.
Kessler, D. and Levitt, S. (1999) “Using Sentence Enhancements to Distinguish Between Deterrence and
Incapacitation”, Journal of Law and Economics, XLII:343-363.
Laband, D.N. and Sophocleus, J.P. (1992) ‘An Estimate of Resource Expenditures on Transfer Activity in the United
States’, Quarterly Journal of Economics, vol. CVII, pp. 959-983.
Lattimore, P.J., Baker, J. and Witte, A.D. (1992) “The Influence of Probability on Risky Choice: A Parametric
Examination”, Journal of Economic Behavior and Organisation, 17:377-400.
Levitt, S. (1996) “The Effect of Prison Population Size on Crime Rates: Evidence from Prison Overcrowding
Litigation” Quarterly Journal of Economics, 111, 319-351.
Levitt, S. (1997) “Using Electoral Cycles in Police Hiring to Estimate the Effect of Police on Crime”, American
Economic Review, 87, 270-290.
Levitt, S. (1998) “Why Do Increased Arrest Rates Appear to Reduce Crime: Deterrence, Incapacitation, or
Measurement Error?”, Economic Inquiry, 36:353-72.
Lindgren, S.A. and Gifford, L.S. (2000), Criminal Justice Expenditure and Employment Extracts Program, NCJ
178273. Washington, D.C., U.S. Department of Justice, Bureau of Justice Statistics.
Marvell, Thomas, and Moody, Carlisle 1996, Specification Problems, Police Levels, and Crime Rates, Criminology
34:609-646.Merritt v. Faulkner, United States Court of Appeals, Seventh Circuit, 1983. 697 F.2d 761.
Mayhew, P. Dijk, J.M. (1997). Criminal Victimization in Eleven Industrialized Countries: Key Findings from the 1996
International Crime Victim’s Survey: The Hague Ministry of Justice WODC. Retrieved on the World Wide Web
on November 2, 2000: http://ruljis.leidenuniv.nl/group/jfcr/www/icvs/data/i_VIC.HTM
Miller, W. (1977), Cops and Bobbies: Police Authority in New York and London, 1830-1870, Chicago: University of
Mirrlees_Black, C. and Ross, A. (1995a) ‘Crime Against Retail Premises in 1993’, Research Findings 26, London:
Mirrlees_Black, C. and Ross, A. (1995b) ‘Crime Against Manufacturing Premises in 1993’, Research Findings 27,
London: Home Office.
Moriyama, T. (2000), ‘Japan’,
gov/bjs/pub/ascii/wfbcjger.txt, accessed July 8, 2000.
Nagin, D. (1998) “Criminal Deterrence Research: A Review of the Evidence and a Research Agenda for the Outset of
the 21 Century” in Tonry, M. (ed.) Crime and Justice: An Annual Review of Research Vol.23, pp.1-42. University
of Chicago Press.
Nagin, D, and Paternoster (1991) “The Preventive Effect of the Perceived Risk of Arrest”, Criminology, 29:561-73.
Nalla, M. World Factbook of Criminal Justice Systems: Singapore. Washington DC: Bureau of Justice Statistics.
Retrieved November 3, 2000, from the World Wide Web: http://www.ojp.usdoj.gov/bjs/pub/ascii/wfbcjsin.txt.
Newman, Graeme (ed) 1999, Global Report on Crime and Justice, New York: Oxford University Press.
Philipson, T.J. and Posner, R.A. (1996), ‘The Economic Epidemiology of Crime’, Journal of Law and Economics, vol.
Polinsky, A. M and Shavell, S. (1999), ‘The Economic Theory of Public Enforcement of Law’, NBER Working Paper
Posner, R. (1992), Economic Analysis of Law, 4th edition, Boston: Little Brown and Company.
Sah, R. (1991) “Social Osmosis and Patterns of Crime”, Journal of Political Economy, 99:1272-95.
of Criminal Justice Systems,
Sampson, R. and Groves, W. (1989), “Community Structure and Crime: Testing Social-Disorganisation Theory”,
American Journal of Sociology, 94, pp.774-802.
Schochet, P.Z., Burghardt, J. and Glazerman, S. (2000), National Job Corps Study: The Short-Term Impacts of Job
Corps on Participants’ Employment and Related Outcomes, Mathematica Policy Research, Inc.
Shapland, J. (2000), “Auditing Criminal Justice” in Fielding, N.G., Clarke, A. and Witt, R (eds) The Economic
Dimensions of Crime, pp. 238-250. MacMillan Press and St. Martin’s Press, London, New York, 2000.
Shavell, S. (1987) “The Model of Optimal Incapacitation”, American Economic Review: Papers and Proceedings,77:
Shaw, S. (ed) (1994), Privatisation and Market Testing in the Prison Service, Prison Reforem Trust, London.
Sherman. L.W., Gottfredson, D. Mackenzie, D., Eck, J., Reuter, P. and Bushway, S. (1997), Preventing Crime: What
Works, What Doesn’t, What’s Promising, U.S. Department of Justice, Office of Justice Programs, Washington, DC.
Smith, S. and DeFrances, C. (1996), Indigent Defense, February, 1996, NCJ-158909, Washington: Bureau of Justice
Stigler, G. (1970), ‘The Optimum Enforcement of Laws’, Journal of Political Economy, vol. 78, pp. 526-536.
Tauchen, H., Witte, A.D. and Griesinger, H. (1994), “Criminal Deterrence: Revisiting the Issues with a Birth Cohort”,
Review of Economics and Statistics, 76, 399-412.
Thomas, C. W. (2000) Home Page accessed 6 August 2000 http://web.crim.ufl/pcp/
Trumbull, W. (1989), “Estimation of the Economic Model of Crime Using Aggregate and Individual Data”, Southern
Economic Journal, 94, 423-39.
United States v. Bergman, United States District Court, Southern District of New York, 1976, 416 F.Supp.496.
van Dijk, F and de Waard, J. (2000), Legal Infrastructure of the Netherlands in International Perspective, Ministry of
Justice, The Netherlands. (http://www.minjust.nl:8080/c_actual/rapport/index.htm).
Walker, I and Sansfacon, D. (2000), Investing Wisely in Crime Prevention – International Experiences, U.S.
Department of Justice, Office of Justice Programs, Washington DC.
Williams, J. Ph.D. Thesis, 1995.
Williams, J. and Sickles, R. (1999), “Turning from Crime: A Dynamic Perspective”, mimeo.
Witt, R. and Witte, A. D. (2000) “Crime, Prison and Female Labor Supply”, Journal of Quantitative Criminology, 16,
No. 1, March, 69-85.
Witte, A. D. (1980) “Estimating the Economic Model of Crime with Individual Data”, Quarterly Journal of Economics
Rule of Law
Source: Kaufmann, Daniel, Aart Kraay and Pablo Zoido-Lobaton, 'Aggregate Govenance Indicators Data and Charts',
http://www.worldbank.org/wbi/governance/gov_matrs_new.htm, accessed June 28, 2000.
Source: Kaufmann, Daniel, Aart Kraay and Pablo Zoido-Lobaton, 'Aggregate Govenance Indicators Data and Charts',
http://www.worldbank.org/wbi/governance/gov_matrs_new.htm, accessed June 28, 2000.
49 Download full-text
Political Instability and Violence 1997-1998
Source: Kaufmann, Daniel, Aart Kraay and Pablo Zoido-Lobaton, 'Aggregate Govenance Indicators Data and Charts',
http://www.worldbank.org/wbi/governance/gov_matrs_new.htm, accessed June 28, 2000.