THE THEORY OF CRIME PREVENTION THROUGH
Ronald V. Clarke (Rutgers)
The theory of crime prevention through environmental design is based on
one simple idea — that crime results partly from the opportunities presented
by physical environment. This being the case it should be possible to alter
the physical environment so that crime is less likely to occur.
Simple, almost obvious as this may be, the idea that crime is partly caused
by physical environment has been a controversial one among criminologists.
In my presentation I will argue that the controversy is based on a
misunderstanding about causes and will review what is now a large body of
evidence that the physical environment plays an important part in crime. I
will then review current approaches to changing the physical environment
and the “opportunity” theories on which these rest. I will also consider the
main criticism of environmental crime prevention that it does not reduce
crime, but merely displaces it to some other time, place or target. Finally, I
will draw out the main lessons from research about successful ways to
implement crime prevention through environmental design.
Three approaches to crime prevention through environmental design
There are three distinct approaches or theories that come under the general
heading of “crime prevention through environmental design”. The term was
coined by the criminologist, C. Ray Jeffery, who published a book in 1971
arguing that sociologists had considerably overstated the social causes of
crime, such as relative deprivation and subcultural influences, and had
neglected both biological and environmental determinants. He went on to
make the general argument that prevention ought to be focused on factors
related to the biology of crime, such as exposure to lead which he thought
caused brain damage and delinquency in children, and to reducing the
environmental opportunities for crime.
1His book met with either indifference or considerable hostility from
criminologists, who were particularly offended by the biological arguments.
The book contained few prescriptions for reducing opportunities, but his
followers, in particular, Tim Crowe (1991), have now developed a
comprehensive set of guidelines to reducing opportunities for crime in the
built environment, intended to guide police, town planners and architects.
These guidelines have been promulgated in hundreds of training sessions
given by Crowe and others throughout the United States, and the approach is
well known by its acronym, CPTED.
The second approach falling under crime prevention through environmental
design is the “defensible space” theory of the architect, Oscar Newman
(1972), who published his famous critique of American public housing at
about the same time as C Ray Jeffery’s book. Newman put much of the
blame for the high crime rates of public housing “projects” on their lay-out
and design. In particular, he criticized the huge, inhuman scale of the
developments, their stark design that made it seem that no-one cared about
them, their location in high crime neighborhoods, and the large buildings
that made it difficult for residents to know who were other residents and who
were intruders. These factors conspired to attract criminal predators who
could commit their crimes with little fear of arrest. Newman, and some other
enlightened architects such as Richard Gardiner, put forward a wide range of
detailed design suggestions to change these conditions and make housing
safer. The purpose of Newman’s suggestions was to encourage natural
territorial behavior on the part of residents by enabling them to give
surveillance to the public areas around their individual residences.
His thesis was savagely criticized by criminologists and other social
scientists, who accused him of “environmental determinism” and of making
simplistic extrapolations to human behavior from the territorial behavior of
animals. Nevertheless, Newman has had an enormous impact on the design
of public housing in many parts of the world. The wholesale abandonment of
tower block buildings for public housing owes much to his arguments about
their criminogenic potential. In recent years, the federal government in the
United States has once again begun to pay attention to Newman (Cisneros,
1996), commissioning him to publish a restatement and defense of his views
The third environmental design approach is situational crime prevention,
which was developed by the British government’s criminological research
2department in the mid-1970’s (Mayhew et al., 1976; Clarke and Mayhew,
1980). Unlike CPTED, and “defensible space”, this approach is not
concerned principally with architectural design and the built environment.
Nor is it focused mainly upon predatory offenses of robbery or burglary.
Rather, it is a general approach to reducing the opportunities for any kind of
crime, occurring in any kind of setting, including airline hijackings, welfare
frauds, obscene phone calls, pub violence and domestic violence, as well as
the conventional predatory offenses.
Most of the research into the relationship between crime and environmental
opportunities has been conducted under the rubric of situational crime
prevention and most of the current “opportunity” theories that now underpin
crime prevention through environmental design, and that have made this a
respectable approach among a large and influential body of “environmental
criminologists”, has been developed by researchers associated with
situational crime prevention.
So if crime prevention through environmental design is now more
theoretically respectable, why was it rejected at first by criminologists and
what has happened to change this state of affairs? I have already mentioned
that Jeffery’s and Newman’s ideas did not accord with those held by most
American criminologists. In America, criminology is principally an offshoot
of sociology and, not surprisingly, social factors are generally held to be the
most important in causation. Oscar Newman, was not even a social scientist,
but an architect! He seemed not only to be ignorant of the findings of
criminology, but had also made some basic statistical errors in his research.
These criticisms, together with some disappointing results from early crime
CPTED experiments made it easy for both Jeffery’s and Newman’s ideas to
be ignored by academics. Even the US federal government lost interest in
the ideas for a period of nearly fifteen years between the beginning of the
1980s and the mid-1990s.
Several developments took place during these years which have greatly
changed the situation:
1. much more evidence has accumulated of the important role of
opportunity in crime;
2. new criminological theories have been developed, including the rational
choice approach and routine activity theory, which give a much greater
role to environmental determinants of crime;
33. a large number of case studies has been published showing substantial
reductions in crime following environmental change;
4. much less displacement of crime has been found as a result of
environmental changes than critics claimed.
The role of opportunity in crime
I will review each of these developments in turn, beginning with the role of
opportunity in crime which I will illustrate briefly with results from three
important studies. The first is not a study of crime but of suicide. I mention it
because, like many serious crimes, suicide is usually seen to be a deeply
motivated act, only committed by very unhappy or disturbed people.
However, a strong and surprising opportunity component appears in suicide
trends in England and Wales during the 1960s and 1970s (Clarke and
From the table (at the end of this article), you should be able to see that in
1958 almost 50 percent of the nearly 5,300 people who killed themselves in
England and Wales did so by domestic gas. This gas contained high levels of
carbon monoxide and was very poisonous. People would kill themselves by
putting their head in the gas oven or by lying down by the gas fire, having
blocked up any gaps under doors or around windows.
During the 1960s, domestic gas began to be manufactured from oil rather
than from coal. As a result, it became less poisonous and the number of
people killing themselves with gas began to decline. In 1968, manufactured
gas began to be replaced by natural gas from the recently discovered North
Sea fields. Natural gas is free of carbon monoxide and is almost impossible
to use for suicide. By the mid-1970s when natural gas had been introduced
throughout most of the country, less than one percent of suicides were by
domestic gas, compared with about 50 percent at the beginning of the
What is deeply surprising is that suicides did not displace wholesale to other
methods. Between 1968 and 1975, total suicides dropped by about one third
from nearly 5, 300 to nearly 3, 700. This was during a time of much
economic uncertainty when one might have expected suicides to increase
and, indeed, was generally increasing in other European countries.
4Why did people not turn to other methods instead? Why did they not
overdose on sleeping pills, shoot or hang themselves, jump out of high
buildings, or put their heads on the railway tracks? The reason is that all
these methods have disadvantages not possessed by gas. It is difficult to
collect together enough pills to kill oneself and many people who take an
overdose have their stomachs pumped and their lives saved. Few people
have guns, and in any case these result in blood and disfigurement. Hanging
oneself or jumping out of a tall building requires courage and resolution.
Lethal domestic gas, on the other hand, used to be piped into most people’s
homes, required little preparation and involved no pain. It is easy to
understand why it was the favored method of suicide in Britain for so long.
Nor is it so surprising that when the opportunity to use it was removed, the
overall suicide rate declined.
The second study concerns motorcycle theft in Germany during the 1980’s
(Mayhew et al., 1989). You will see from the first column of the table that
thefts of motorbikes had drastically declined from about 150,000 in 1980 to
about 50,000 in 1986. If you who do not already know the reason, you will
be most surprised to learn it. It was because in 1980 it was made illegal to
ride a motorbike in Germany without a helmet. This meant that anyone
stealing a motorcycle without a helmet would be readily spotted. The law
was gradually enforced more strictly during the period and resulted in the
large decline in motorcycle thefts. This suggest a much stronger
opportunistic component in motorcycle theft than anyone would have
The second and third columns of the table show the totals for car and bike
thefts during the same years. These provide some limited evidence of
displacement because thefts of cars went up from about 64,000 to 70,000.
Thefts of bicycles also increased between 1980 and 1983, but by the end of
the period had declined again below the numbers for 1980. Altogether, it is
clear that at best only a small proportion of the 100,000 motorbike thefts
saved by the helmet laws were displaced to thefts of other vehicles.
Again, a little thought shows why this may not be surprising. Motorbikes
may be particularly attractive to steal. Young men who comprise most of the
thieves find them much more fun to ride than bikes. Even if the intention is
only to get home late at night, it is much easier to ride a motorbike for a few
miles than a bicycle. Motorbikes may also be easier to steal than cars since
the latter have to be broken into before they can be started.
5The third study comes from the United States and concerns the dramatic
increase in residential burglary during the 1960s and 1970s. A careful
analysis by Lawrence Cohen and Marcus Felson (1979) showed that this
increase could readily be explained by a combination of two changes that
had occurred in society during this period. Together these had greatly
increased the temptation and opportunity for burglary. Temptation had been
increased by the vast increase in light-weight electronic goods such as TV’s
and VCR’s in people’s homes that could readily be sold. The opportunity to
commit burglary was greatly increased as a result of far more women going
out to work. This meant that far more homes were empty during the day and
therefore unguarded. The combination of more attractive goods to steal and
more unguarded homes easily accounted for the increase of burglary
without having to assume any increase in the criminality of the population.
Studies like these have helped to bring about a change in criminological
theorizing. In particular, it is now more widely recognized that most
traditional criminological theories are really theories of criminality or
delinquency and not theories of crime. Traditional theories deal in the
factors that cause people or groups to be disposed to criminal action. In other
words, they deal in behavioral tendencies, not behavior itself.
The point can be clarified by revisiting a truism of psychology, which is that
behavior is the product of the interaction between the person and the setting.
To put this in the language of criminology is to say that crime (a behavior) is
a product of the interaction between a criminal or delinquent propensity and
a criminal opportunity. While it is necessary to understand the factors that
result in criminal dispositions, this alone does not explain the occurrence of
crime. To explain crime, one must also explain the interaction between
disposition and opportunity. This is what most traditional criminology has
failed to do. Instead, it has assumed that explaining criminal dispositions is
the same as explaining crime.
This error is now being corrected. Criminologists are beginning to take an
interest in developing, not just theories of criminality or delinquency, but
what are called “theories of crime” (or, sometimes, “opportunity” theories).
The two best known of these are routine activity theory and the rational
choice perspective. Though complimentary, these both deal with somewhat
6different questions. Routine activity theory (Cohen and Felson, 1979) seeks
to explain how changes in society in the numbers of “suitable targets” for
crime, or in the numbers of “capable guardians” against crime results can
lead to more or less crime. One example of the application of this theory is
Cohen and Felson’s explanation, that I mentioned, for the increased burglary
rate in the United States in the 1970s. This was due to the increase of
“suitable targets” in homes (VCRs and TVs) and the decline of “capable
guardians” (women no longer at home in the day).
The rational choice perspective (Cornish and Clarke, 1986) tries to
understand crime from the perspective of the offender. It asks: What is the
offender seeking by committing crime? How do offenders decide to commit
particular crimes? How do they weigh the risks and rewards involved in
these crimes? How do they set about committing them? If prevented from
committing them, what other crimes might they choose to commit?
It should be clear from these questions that the rational choice perspective is
directly concerned with the thinking processes of offenders, how they
evaluate criminal opportunities and why they decide to do one thing rather
than another, Indeed, why they choose to obtain their ends by criminal and
not legal means.
This perspective has helped to explain why displacement does not always
occur (think about the reasons I gave for suicides not switching to other
methods and why motorcycle thieves did not begin to steal cars or bicycles).
It has also been helpful in thinking about different ways to reduce
opportunities for crime. Sixteen different ways to reduce opportunities for
crime are described in situational crime prevention falling under four
objectives, taken from the rational choice perspective. These are:
1. to increase the perceived difficulty of crime;
2. to increase the perceived risks of crime;
3. to reduce the anticipated rewards of crime;
4. to remove excuses for crime.
More than eighty examples of success using these opportunity-reducing
measures have been documented in careful research studies. Twenty three of
these studies are reprinted in my book, “Situational crime prevention:
Successful case studies” (Clarke, 1997), but I can only list a handful of these
now. They are listed in the table under the four principal ways to reduce
7opportunities for crime. You will see that some of these measures were
devised here in Sweden.
I have mentioned “success”, but what about displacement? The gas suicide
and motorcycle theft examples have already shown that displacement is not
inevitable. But what happened as the result of the measures that I listed in
my table? For example, did the check forgers frustrated by the new check
cashing rules in Sweden turn to some other kind of fraud or theft? Did the
post office robbers thwarted by the anti-bandit screens in London begin to
take customers hostage or attack the armored vehicles collecting money
from the post offices? Were criminals shifted by the improved street lighting
in the council housing estate in Dudley to another nearby estate where the
lighting was not improved? And so forth.
In truth, it is difficult to make definitive studies of all these possibilities, but
many published evaluations include some investigation of displacement.
Four years ago, Rene Hesseling (1994) reviewed the literature on
displacement for the Dutch Ministry of Justice. He examined 55 studies in
which displacement had been studied. What he found was very revealing. In
22 of the studies, no evidence of displacement was found. In the remaining
33 studies some evidence was found, but the displacement was often
relatively small. In no case, did the crime displaced elsewhere equal the
In addition, researchers are now beginning to find evidence of the reverse of
displacement so that focused situational prevention projects have produced
wider reductions in crime beyond the direct reach of the prevention
measures. A clear example comes from a university campus in Southern
England where CCTV cameras were introduced to monitor the car parks
(Poyner, 1997). Car theft fell not only in the three car parks given
surveillance by the cameras, but also in a fourth car park that the cameras
could not monitor because the view was obstructed by buildings. It seems
that potential car thieves were unsure about the coverage of the cameras and
decided to avoid the university altogether. This phenomenon, whereby the
preventive benefits spread more widely, has been called the “diffusion of
benefits” and is now being found regularly in a variety of studies. It greatly
enhances the appeal of crime prevention through environmental design.
8Summary and recommendations
Before concluding with some recommendations about applying crime
prevention through environmental design, let me summarize my presentation
so far. I have argued that the earlier claims made for crime prevention
through environmental design were rejected by criminologists who felt that
the physical environment plays a relatively unimportant part in crime
causation. Much of this criticism was misplaced because criminologists had
generally not recognized that their theories were mostly concerned with the
development of criminal dispositions, not the occurrence of crime.
Growing evidence about the role of opportunity in crime has encouraged the
development of some new “opportunity” theories, including routine activity
theory and the rational choice perspective. The latter in particular has
provided the theoretical basis for situational crime prevention, which has
now accumulated a solid record of crime reductions achieved in many
different contexts. Displacement has not proved such a threat as once
thought and there is growing evidence of diffusion of benefits from crime
There is no doubt that crime prevention through environmental design will
be increasingly be used to help protect society from crime. Indeed, it already
is an integral part of crime control policy in several European countries. But
what has been learned in the last 25 years about its application to town
planning and architectural design, which is of particular interest to many of
you at this meeting? Let me conclude with what I believe are the main
1. Reducing access and enhancing territoriality and surveillance, as
advocated by Oscar Newman, and in Tim Crowe’s more recent guidelines,
represent a somewhat narrow crime prevention approach. A much wider
range of opportunity-reducing techniques has been identified in recent
situational crime prevention research.
2. The image of the “predator from outside”, which underlies the original
defensible space thesis (and Crowe’s guidelines), ignores the fact that much
crime is committed by residents and other legitimate users of the space.
Access controls and improved surveillance may be of limited value in
preventing crime by residents.
93. Opportunity-reducing measures appropriate to certain offenses may not be
appropriate to others. This means that crime prevention through
environmental design must be tailored to the specific problems occurring in
particular settings. Standard packages of measures will not usually reduce
4. In any setting, there are always choices to be made in the opportunityreducing measures to be implemented. These choices must consider the
practicality and sustainability of the various options, their social
acceptability and their economic costs. It is nearly always possible to
identify practical, low cost, unobjectionable ways of reducing opportunities
5. Numerous evaluations have shown that, by themselves, design changes
can bring about only rather modest reductions in crime. If substantial
reductions are to be achieved, design improvements usually have to be
supported by changes in management or policing of the built environment.
6. When crime prevention design changes are being made to existing
environments, care must be taken to ensure that residents and other
legitimate users are consulted in detail about the proposed improvements.
Otherwise, changes might be resisted and prove impossible to implement as
So long as crime prevention through environmental design is applied
carefully and intelligently, with due regard to these and other lessons of past
practice, there is every reason to believe that it can achieve worthwhile
reductions in crime, wherever it is employed.