Read an Excerpt
CHAPTER ONE
What Americans Fear
By longstanding habit, Americans use the terms "crime" and "violence"
interchangeably. When expressing concern about urban conditions we
commonly talk about "the crime problem" or "the violence problem" as if
they were the same thing. When drive-by shootings create newspaper
headlines, we demand that our elected officials do something about crime.
At the core of this interchangeable usage of crime and violence is the
belief that crime and lethal violence are two aspects of the same problem.
It is widely believed that there is much more crime of all kinds in the
United States than in other developed countries. With so much crime and
so many criminals, the high rates of lethal violence in the United States
seem all but inevitable to many observers.
The mission of this chapter is to demonstrate that rates of crime are not
greatly different in the United States from those in other developed
nations and that our extremely high rates of lethal violence are a separate
phenomenon, a distinct social problem that is the real source of fear and
anger in American life. The chapter will provide a review of some
objective conditions of crime and lethal violence in the United States and
link these empirical patterns to the subjective dimensions of fear in urban
America. The first section of this chapter will provide data on crime and
lethal violence in international perspective. This section provides a
preview of the extensive research findings that will be presented in
Chapters 2 to 5.
The second section of this chapter presents a simple theory of why
lethal violence is a major concern in developed nations as well as a
problem of great importance in the United States. The third part of the
chapter considers some complicating factors in the relationship between
crime, violence, and citizen fears. We identify a process of categorical
contagion that leads citizens to fear lethal violence in a broader variety of
settings than those that carry any substantial risk to life and limb. To live
in an environment where robbery presents a serious hazard to its victims
seems to provoke citizens to fear for their lives even from particular forms
of crime that do not place their victims at mortal risk.
The final section of this chapter shows that general crackdowns on
crime are inefficient and potentially self-defeating methods of reducing
the risks of lethal violence. It thus appears that one natural reaction to
lethal violence, a heightened fear of all kinds of crime, may lead to
ineffective responses to the life-threatening behavior that is the core
concern of the fearful public. Citizen fears may systematically point
public policy toward crime in the wrong direction.
Crime and Violence in International Perspective
Los Angeles and Sydney
Los Angeles is the second largest city in the United States with a 1992
population estimated at 3.6 million for its crime statistics reporting unit. It
is a multiracial, multicultural city on the Pacific Coast with a crime rate
that by most accounts is its most serious civic problem. Sydney,
Australia, is also a city of 3.6 million located on the Pacific Coast of the
continent. While multicultural by Australian standards, the ethnic and
language mixture falls far short of that in Los Angeles. Crime in Sydney is
a serious annoyance, but not a major threat to the continued viability of
the city or to the health and welfare of its citizens.
Figure 1.1 compares the volume of four crimes reported by the police
in Sydney and Los Angeles by expressing the number of offenses in
Sydney as a percentage of the Los Angeles crime volume. Since the
population of the two cities is the same, the crime volume comparison is
also a crime rate comparison. The theft category reported at the far left of
Figure 1.1 includes most forms of stealing that are unaggravated by
elements that the law regards as increasing the gravity of the offense.
This is the most common offense reported in all cities, and the two
jurisdictions under review are no exception. Sydney reports just over
90,000 theft incidents, roughly three-quarters of the volume reported in
Los Angeles.
Burglary is an aggravated form of theft where the offender breaks and
enters private property in order to steal. The crime statistics for Sydney
report the offense under two headings: breaking and entering a dwelling,
and breaking and entering a building that is not a dwelling. The volume
of such crimes in Sydney during 1992 exceeded 63,000, about 10 percent
more than the number of burglaries reported in Los Angeles.
The pattern noted for burglary contrasts sharply with robbery, the
other major category of aggravated property crime. Robbery is defined as
the taking of property from the person of another by force or by the
threat of force. In 1992 Los Angeles reported 39,508 robberies while
Sydney reported 4,942, one-eighth the Los Angeles rate. The ratio of
burglaries to robberies in Los Angeles is just under 3 to 2; the ratio of
burglaries to robberies in Sydney is greater than 12 to 1.
The final crime category reported in Figure 1.1 is for homicides
resulting from intentional injury. There were fifty-three such offenses
reported by the police in Sydney during 1992, a crime volume equal to 5
percent of the 1,094 homicides reported by the Los Angeles police that
same year. The difference between the two cities in rates of criminal
homicide exceeds an order of magnitude. The citizens of Sydney can thus
live with their high crime rate in relative comfort because they are not
dying from it in large numbers.
The major statistical conclusion one draws from an inspection of
Figure 1.1 is that the nature of the comparison between Sydney and Los
Angeles depends on what is being compared. For theft and burglary the
two cities are quite similar. For robbery and homicide they are vastly
dissimilar.
The substantive conclusion to be drawn from this statistical pattern
can be stated in two ways. It seems beyond dispute that what separates
the two cities is not the amount of crime they experience, but the
character of the crime they experience. But the point can be put more
sharply than that: What is distinctive and threatening in Los Angeles is
not a crime problem but a violence problem.
New York City and London
New York City is the largest city in the United States with a population at
7 million. London has a city population of 6.6 million. Figure 1.2 shows
London crime rates per 100,000 using New York City rates as a standard
for comparison. The statistical comparison in Figure 1.2 is even more
surprising than that concerning Los Angeles and Sydney, and to the
same effect. London has more theft than New York City and a rate of
burglary 57 percent higher. But the robbery rate in London is less than
one-fifth the robbery rate in New York City and the homicide rate in
London is less than one-tenth the New York City figure.
The total number of offenses per 100,000 citizens in Figure 1.2 is higher
in London than in New York City. If total crime rates were the problem,
Londoners should live in fear or New Yorkers in relative complacency.
They have the same magnitude of crime. But with death rates eleven
times as high as London, the population of New York City is far from
comfortable. Lethal violence is New York City's distinctive problem, not
crime, and lower rates of general theft are no consolation for huge death
toll differences.
If readers are wondering whether these data are selective and
misleading crime rates, or are numbers that are produced by the
peculiarities of different reporting systems, the next set of comparisons
should provide some reassurance that our city comparisons are part of a
broad and consistent pattern.
Twenty Countries
We can show clearly that America's special problem is violence and not
crime by comparing the results of a twenty-nation survey of citizens
about the rate at which they were victims of crime with World Health
Organization data on death from assaults for the same nations. Figure 1.3
(see p. 8) shows the violent death rates for each group of five nations
with the highest crime levels, then the next two highest, and finally the
lowest crime categories.
There are several indications that a country's crime rate is substantially
independent of its rate of lethal violence. First, the variation in violent
death rate is quite large within the separate crime rate categories. Within
the group of highest crime nations, the homicide rates vary by a factor of
eleven, in the next group by a factor of five, in the third group by a factor
of three, and in the lowest crime group by a factor of eight. In contrast,
the median homicide rates for the four different crime categories are
clustered between 1.3 and 2.2. So knowing which crime rate category a
country belongs in does not tell one anything much about what rate of
violent death that country suffers.
And knowing a country's violent death rate does not predict much
about its crime rate. The lowest death rate country (England) has a crime
rate just over average. The next lowest violence nation is Japan, which
has the lowest crime rate also. The third lowest death rate country is the
Netherlands, in the highest crime rate group. The pattern is just as
opaque at the top of the violence distribution. The most violent country,
the United States, has a high crime rate as well. The next most violent
country, Northern Ireland, is in the lowest crime rate group.
This data set provides a multinational example of the central point that
lethal violence is the crucial problem in the United States. It shows the
United States clustered with other industrial countries in crime rate, but
head and shoulders above the rest in violent death.
It also suggests that lethal violence might be the best predictor of
citizen fear on a transnational basis. Where would you rather live when
examining the map of Figure 1.3? In England with its high crime rate and 0.5
deaths per 100,000 or in Northern Ireland with a much lower crime rate
and nine times the death rate? Your money or your life?
Judging from Figure 1.3, the United States has about the same rate of
crime and prevalence of criminality as the Netherlands and Australia. But
ours is by far the most dangerous country to live in. We currently have a
Netherlands-size crime problem and a king-size violence problem that
threatens the social organization of our cities. Which problem is at the
root of citizen fear in the United States? Which problem should we try to
solve?
Three Dimensions of Fear
What types of loss from crime produce fear and anxiety among citizens?
A common-sense calculus would emphasize three aspects of the types of
criminal harms that are of the largest significance in provoking fear:
1. the importance of the interest threatened
2. the capacity of insurance and compensation schemes to restore the
welfare of crime victims, and
3. the extent to which potential crime victims feel they can control
their risks of loss by altering behavior.
For a mixture of obvious and nonobvious reasons, lethal violence is
the most frightening threat in every modern industrial nation. The
obvious reason why lethal violence is a priority concern is the
importance people attach to personal survival. Physical survival, security
from assault, freedom from the pain inflicted by serious intentional
injuries are among the most basic interests citizens have. On this basis
alone, we would expect life-threatening assault to be at the top of any
citizen's fears about crime.
But the effectiveness of insurance and compensation schemes to ease
the pain of property crime when compared with the impossibility of
restoring life and health is an additional reason why serious violence is
the central reason citizens fear predation. The security of property from
criminal trespass can be achieved in the modern state by creating safe
methods of preserving property interests and by facilitating insurance
and other loss-spreading mechanisms to function when property is taken.
Most economic assets cannot be taken by a thief in the age of the bank
account and real estate title. Those chattels that remain at risk--the
automobile, the home computer, and the bicycle--are usually the subject
of automobile or household insurance. The percentage of gross national
product that is redistributed by larceny and burglary in developed
countries is quite small. When one of us asked a group of students,
"What happens when someone steals your new BMW?" one student
replied, "You get a better BMW," an arch response that nonetheless
reflects social perceptions about automobile insurance with some
accuracy. For most citizens, replacing one BMW with another comes quite
close to making the crime victim whole; auto theft is amenable to a
process we would call commensurate compensation.
By contrast, no amount of life insurance will give the insured back his
life if he is killed in a street robbery. So even when programs are in place
to deal with the financial costs of serious violence, the lack of
commensurate compensation will make those criminal harms that cannot
be reversed more worrisome to potential victims. In this sense, one
natural byproduct of extensive insurance and property security is that
citizens will concentrate their crime anxieties on those losses that cannot
be effectively redressed. The special priority accorded to threats like
homicide, forcible rape, and life-endangering injury will be more apparent
in a system where most citizens have little to fear from standard forms of
property crime.
The third dimension that we believe influences the level of fear
expressed about particular criminal harms is the degree to which potential
victims feel they can control the risks of becoming victims. The prospect
of what the newspapers call "random violence" is particularly problematic
because it suggests that there is no set of precautions a potential victim
can take to reduce the personal risk of victimization. The more members of
the public feel the risks of a particular harm are within their control, the
more secure citizens will feel about their personal situation. But if
potential victims feel there is no way they can take action to modify risks,
this will heighten their anxiety about a particular risk. Later in the book,
we will suggest that one reason citizens fear stranger violence more than
they fear being killed by friends or family is because they feel more
control over their choice of personal acquaintances, while the strangers
they encounter are not as easy to choose or to reject.
If these three dimensions of risk are the significant factors in
predicting citizen fear, two points of general applicability can be made
about lethal violence as a crime control priority in developed nations.
First, there is good reason to suppose that life-threatening violence is the
most feared aspect of the threat of crime, even in countries where the rate
of homicide is quite low. While an American observer might be amused to
find fear of gunshot wounds is a focal concern in Japan, where there were
thirty-two gun homicides in 1995 (Kristof 1996), the small chance of a
major harm that cannot be effectively compensated or avoided might be
more worrisome than less catastrophic injuries, even if these occur with
much greater regularity. It may thus be rational for all of the countries
covered in Figure 1.3 to hold lethal violence as the number one concern.
A second general point is that a special concern with lethal violence
may be one consequence of successful social and economic
development. The greater the success of a society in producing and
distributing property, the larger the presence of insurance and
compensation, the less average citizens will need to worry about their
automobiles and their bicycles. This allows public concern about crime to
return to bedrock worries about physical security. So the fact that lethal
violence is a social control priority is not always a measure of social
pathology. The greater the success of crime prevention and loss
spreading, the more likely it will be that security of life and limb are the
residual concerns of the citizenry.
So life-threatening violence should be a central concern in all
developed nations, but nowhere in the Western world should lethal
violence generate the level of attention it demands in the United States.
Even if the rate of lethal violence is not relevant to whether such acts are
the chief concern of law enforcement, the sheer amount of lethal violence
is an important determinant of how much public concern is justified. A
homicide rate five times the average of the developed nations and twice
that of Northern Ireland in a time of troubles is an important cause of
concern for the major developed nation in the modern world. Chronically
high rates of lethal violence generate insecurities in the United States that
are qualitatively different from those found in other advanced countries.
Lethal violence is the most serious social control problem in every
developed nation; it is by far a more serious problem in the United States than
in any society with which we would care to compare ourselves.
Crime, Violence, and Citizen Fear
One sharp contrast between Sydney and Los Angeles invites us to
consider the relationship between rates of life-threatening personal
violence and public perceptions of the seriousness of crime as a problem.
As measured by the political importance of crime as an election issue, or
the degree to which fear of crime is mentioned as a major disorganizing
influence on urban life, Sydney does not seem to have a significant crime
problem. To be sure, housebreakers are not popular figures in New South
Wales and complaints about crime are quite common, but terror is not.
Los Angeles is a city where fear of crime and of criminals is arguably
the single most important social and political issue for the majority of
citizens. Scientific surveys of public opinion reflect this difference to
some degree, but they do not capture the palpable difference between
crime as an annoyance in Australia and crime as a fundamental threat to
social life in Los Angeles. The statewide Field polling organization in
California finds crime to be the number one problem reported in the state;
concern about policy toward criminals dominated the 1994 California
elections to an extent that would be unthinkable at any level of
government in Australia. The intensity of citizen concern in California is
also reflected in the demand for substantial changes in criminal justice
policy. The prison population in California grew 400 percent in the fifteen
years prior to the 1994 elections, but California voters nevertheless
overwhelmingly supported a referendum in 1994 that required a
twenty-five years to life sentence for anyone convicted of a third felony if
the offender's previous convictions had been as serious as
housebreaking.
Since general levels of nonviolent crime in Sydney and Los Angeles
are closely similar, why not conclude that it is levels of lethal violence
rather than of crime generally that determine the degree of public fear?
Why else would similar numbers of criminals and rates of crime lead to
such a sharp cleavage in public response? The question is an important
one, but far from easy to answer with confidence. Evidence regarding the
relationship between rates of different types of crime and public attitudes
is surprisingly sparse, and the specific question of the influence on
attitudes of rates of violence, rather than rates of crime generally, has not
been addressed systematically.
There are two different issues involved in determining the relationship
between rates of violence and public fear: the salience of lethal violence
as a focal point for citizen fear, and the influence of rates of violence on
levels of public concern about violence. On the first issue the answer
seems clear: when citizens are afraid of crime it is life-threatening,
personal violence that dominates their attention. On the second issue, the evidence
is far from clear. While the objective risks of violence undoubtedly influence
the level of public fear, so also do many other variables. And the extent
to which differential levels of public concern can be explained by
differences in objective risk is not known.
To characterize concern about serious personal violence as the
dominant image in public fear of crime may seem like an overstatement.
Chapter 4 will document the relatively low levels of danger to life
associated with residential burglary as a crime, but residential burglary is
a crime that citizens greatly fear in California. Indeed, including residential
burglary as a triggering felony in the California "three strikes" sentencing
proposal was vigorously supported by the public even though it tripled
the cost of the program (Zimring 1994). Is this not evidence that
nonviolent threats are as salient to individuals as violent ones?
Probably not. Public fear of burglary is generated by images of the
worst thing that could happen in the course of a housebreaking, rather
that the kind of things that usually do happen when burglars appear. The
majority of burglarized dwellings are unoccupied at the time of the
invasion, but the image of burglary that provokes public fear is of the
burglary of an occupied dwelling. The great majority of burglars would
react nonviolently in any interaction with household members, but the
image of burglary that produces high levels of citizen fear finds the victim
defenseless in bed and at risk of murder. It is the worst-case burglar that
provokes those citizens who express high levels of fear regarding
residential burglary. This is an issue that could be explored by carefully
constructed survey research. To date, however, the question of what
images preoccupy citizens with high levels of fear regarding particular
types of crime has not been investigated.
But fear of the worst-case burglar does not explain the contrast
between Sydney and Los Angeles. Why would housebreakers provoke
much more fear in Los Angeles than in Sydney? One reason might be the
fact that many people have homogeneous images of "the criminal." That
is to say that they do not think of robbers and burglars as different sorts
of people, but rather imagine the criminal offender who threatens their
sense of security as a composite of the personal characteristics of the
criminal offenders they have heard about. If this homogeneity of image
phenomenon is operative, citizens who live in environments where
homicidal attacks are common will fear all kinds of contact with criminal
offenders much more than citizens whose composite image of the criminal
offender derives from a general environment that experiences less lethal
violence.
In an urban environment where armed robbery frequently leads to the
death of victims, the purse snatcher and the burglar will acquire much of
the threatening character of the robber because the composite
generalized image of the criminal that conditions public fear acquires the
characteristics of the lethal armed robber. The fear generated by the
kidnap and murder of Polly Klaas in California provokes long sentences for
residential burglars because the burglar in the citizen's scenario has acquired the
characteristics of Polly Klaas's killer.
High levels of interpersonal violence could thus generate a process
we call categorical contagion. This is the agency whereby citizens come
to fear many forms of criminal behavior because they imagine them all
committed by extremely violent protagonists. Lower general levels of
violence may be associated with less pressure toward categorical
contagion because there is less in the way of frightening violence to
condition the citizen's image of the criminal threat.
Processes of categorical contagion may operate in social life well
beyond the frontiers of the criminal code. Just as personal safety and
bodily security are the core concerns of fear of crime, concern about
vulnerability to assault can produce fear of a wide variety of social
encounters that include rudeness and incivility and even face-to-face
contact with strangers in the streets if those strangers are seen as
threatening. If a person carries profound feelings of physical
vulnerability into a social setting, even ambiguous or innocuous
behavior can produce substantial levels of anxiety.
It is also important to recognize that this process of categorical
contagion may be a two-way street in which a high level of anxiety about
strangers or face-to-face interaction may express itself as a fear of
becoming a victim of a violent crime. Just as substantial anxieties about
being robbed or attacked may make a person apprehensive about
encountering strangers, an intense but nonspecific fear of strangers or
foreigners or black people may produce more specific concerns that the
subjects of our apprehension intend to assault us.
Some Social Causes of Fear
But what social circumstances predict variations in citizen concern about
safety in a modern industrial society? This is a complicated question and
one that has not been squarely addressed in the social science literature
on the fear of crime. We would expect at least three major influences on
the level of fear regarding serious crime: (1) the amount and seriousness
of violent crime, (2) the level of fear-arousing social conditions in the
immediate physical environment of the subject, and (3) the amount and
perceived seriousness of fear-arousing cues in the mass media and the
personal social universe of the subject.
How important are variations in actual risk in the mix of cues that
produce levels of citizen fear of violence? We would expect variations in
the risk of serious violence to be a major determinant of levels of fear that
exert influence in a variety of ways. The higher the rate of serious
violence, the larger the chance that the average citizen will have personal
experience as a crime victim or be in some social relationship to a violence
victim. The larger the risk of serious violence, the stronger the associations
between fear of violence and various fear-arousing cues in the citizen's immediate
social environment. The higher the number of people who get shot,
stabbed, or mugged, the more fear-arousing will be citizen contact with
boom boxes, broken windows, or threatening-looking strangers. Finally,
we would expect that both the number of social cues and the amount of
media attention to violence would be directly influenced by the rate and
seriousness of violence, although we are more confident of this
relationship in the contacts of the citizen than in the quantum of media
cues about violence.
We would expect to find a positive association over time and
cross-sectionally between risks of life-threatening violence and the mix of
cues that determine levels of public fear. But variations in risk are by no
means the only influences on levels of public consciousness in mass
society. Processes of categorical contagion will link levels of public fear
of violence to fluctuations in other social conditions that make people
anxious and insecure. Further, to the extent that the character and
quantity of media attention to violence is a variable that fluctuates
independently of trends in risk, this will also influence levels of fear,
particularly when citizens lack more direct experience. It thus seems
probable that fear of violence in the 1990s is, to some extent, a media
event independent of changes in other social conditions. We would also
expect that variations in media coverage will be of larger importance in
determining levels of concern and fear amongst groups that lack
significant first- and second-hand experience of violence than among
persons with more direct experience.
We do not know the maximum level of public fear that can be
produced and sustained in an environment of general social anxiety, but
low levels of lethal interpersonal violence. An attempt to import an
American-style "law and order" campaign into English electoral politics
would be a natural experiment to see whether fear of lethal violence can
be induced and sustained at high levels without a high death rate from
violence as a feature of the social environment.
A similar issue concerns the generality of fear induced by incivility and
disorder. James Q. Wilson and George Kelling argued persuasively in
their "Broken Windows" article that indications of incivility and disorder
produce citizen fear and the demand for law enforcement and social
control (Wilson and Kelling 1982). One reason such indications could
arouse fear is that they convey to many persons the message that they
are vulnerable to more direct and more violent predation. But will the level
of fear produced by disorder and incivility be as great where rates of
lethal violence are low as they will be where those rates are high? The
context in which the broken-window argument was made was urban
conditions with high rates of lethal violence. The extent to which fear of
lethal violence and fear of concentrated threats to public order feed off
each other should be a priority concern in the social psychology of crime
fears. And the social psychology of citizen fears should be an important
topic for scholarly research.
Crime Policy as Violence Policy
Life-threatening violence is, of course, against the public policy of the
modern state in all but exceptional cases, but so are a wide variety of acts
that range from larceny to illegal drug taking to sexual exploitation of the
immature. What, if anything, is wrong with public policy that treats
violence only as one of the many crime problems that are best addressed
by police, prosecutors, and prisons?
This section outlines some of the difficulties associated with
misdefining American violence as solely a crime problem. In pursuing this
analysis, we do not deny that crime in the United States is destructive,
costly, and disorganizing. Rather, we argue that it is the violent strain in
American social life which causes the special destruction and
disorganization produced by American crime.
