The Great São Paulo Homicide Drop

Abstract

The homicide rates in the city and state of São Paulo were cut in half in the years from 2001 to 2007. The decline in the city of São Paulo was especially striking and parallels the decline in New York City in the 1990s. It can be confirmed with a number of independent data sources and was significantly larger than in other Brazilian cities. The decline may be attributed to more effective policing methods including the better enforcement of strict gun-control legislation. It demonstrates that effective measures can be taken to reduce lethal crime in a developing country without waiting to solve underlying socioeconomic problems.

Full text

The Great São Paulo Homicide Drop
Prepared for the June 2009 meetings of the Latin American Studies Association
Rio de Janeiro, Brazil
forthcoming in Homicide Studies
by Ted Goertzel
Professor of Sociology
Rutgers University
Camden NJ 08102
609 744-7475
goertzel@camden.rutgers.edu
and Tulio Kahn,
Coordinador de Análise e Planejamento
Secretaria de Segurança Pública
Estado de São Paulo
tkahn@sp.gov.br
Forthcoming in Homicide Studies

The Great São Paulo Homicide Drop
Abstract
The homicide rates in the city and state of São Paulo were cut in half in the years
from 2001 to 2007 (SESP, 2008). The decline in the city of São Paulo was especially
striking and parallels the decline in New York City in the 1990s. It can be confirmed
with a number of independent data sources, and was significantly larger than in other
Brazilian cities. The decline may be attributed to more effective policing methods
including the better enforcement of strict gun control legislation. It demonstrates that
effective measures can be taken to reduce lethal crime in a developing country without
waiting to solve underlying socioeconomic problems.
2

With an estimated population of 11,000,000 within the city limits, and with
approximately 20,300,000 in the metropolitan area (IBGE, 2007), the greater São Paulo
metropolitan area is by one estimate the seventh largest urban agglomeration in the world
(Brinkhoff, 2006). The metropolitan area dominates the state of São Paulo with an
estimated population of 41,000,000, about the size of Argentina. The remarkable
homicide drop in Sao Paulo in the first years of this century is as striking and important
as the much better known homicide drop in New York City in the 1990s, but it has not
been as extensively reported or analyzed.
Before the recent homicide drop, Brazil’s high homicide rates were frequently
attributed to the country’s high levels of poverty and inequality. In a recent book,
historian Luís Mir (2004) insisted that Brazil was in a state of civil war and characterized
São Paulo and Rio de Janeiro as metropolises of death. Mir insisted that “nothing can be
done about the problem until the majority and the minority sit down and discuss the slices
of the pie” that each social class receives (Geração Online, 2004). But his book was
published three years after the homicide rates had begun a sharp decline despite the fact
that no such radical re-slicing of the socioeconomic pie had taken place. This parallels
the experience of leading American criminologists James Q. Wilson (1995) and John
DiIulio (1996) who published works in the early 1990s predicting massive increases in
crime rates after the rates had already begun a precipitous decline.
These analysts erred by attributing cyclical peaks in crime waves to persistent
social and economic problems, and by underestimating the extent to which violent crime
has its own dynamics and can be treated as a separate problem. When crime waves get
3

out of hand the public demands action, political leaders allocate more resources, and the
criminal justice system does its best to respond. In both Brazil and the United States,
many police authorities took effective action to reduce violent crime without waiting for
underlying social problems to be resolved. This was also true in Colombia where
homicide declined 15% in the three years from 2003 to 2006 (Casa de Nariño, 2006). By
contrast, homicide rates in Venezuela increased 67% from 1999 to 2005 (Romero, 2006)
despite a booming economy and a populist government that strove to redistribute wealth
to the poor.
Research on socioeconomic factors in crime in Brazil has found complex and
varied patterns not easily reduced to generalities about inequality and oppression
(Andrade and Lisboa, 2000; Mendonça et al, 2003; Drummond, 2002; Cerqueira and
Lobão, 2003a; Cerqueira and Lobão, 2003b; World Bank, 2006). Homicide rates are not
always highest in the poorest neighborhoods or at the times of highest unemployment.
Young men with low incomes are the most frequent victims as well as perpetrators of
violent crime. Most homicides do not involve individuals of markedly different social
standing.
Studies by Coelho (1988) and Paixão (1988) in the state of Minas Gerais found
that socioeconomic factors were less important than the efficiency of the criminal justice
system in explaining variations in homicide rates. In a cross-sectional study in
metropolitan São Paulo, using data from 1970 to 1984, Pezzin (1986) found that poverty,
unemployment and urban density correlated with property crimes but not with crimes
against persons. Sapori and Wanderly (2001) tried to establish a relationship between
unemployment and homicide rates in Rio de Janeiro, São Paulo, Belo Horizonte and
4

Porto Alegre but were unable to establish robust results. A number of more complex
econometric models have offered contradictory findings (Crequeira and Lobão, 2003a).
A study of homicides by census tract in the city of São Paulo from 2000 to 2003
(Kilsztajn, et al. 2005) found that, although most of the victims were men with low
incomes, poverty did not explain the variation in homicide rates between neighborhoods.
There were poor census tracts with high homicide rates and poor census tracts with low
homicide rates. The critical difference seemed to be the presence of organized drug
trafficking, although the data on drug trafficking patterns are less reliable. Similar
conclusions were reached by Beato Filho et al (2001: 1170) in Belo Horizonte and by
Baierl (2004: 145) in the city of Santo André in the industrial suburbs of São Paulo.
This paper examines trends in homicide and related crimes in the city and state of
São Paulo, Brazil’s largest population center, and puts these trends in the context of
trends elsewhere in Brazil. It draws on data from a number of different sources that offer
insight into the causal factors involved. Finally, it places São Paulo in the context of the
emerging literature on crime declines around the world (Zimring, 2007).
Trends in Homicide and in São Paulo. Chart One shows the homicide rate in
the city of São Paulo for the years from 1980 to the first half of 2008, a period long
enough to show a full historical cycle. These rates are drawn from two different sources.
The first source, from health department records, includes all homicides, including
negligent homicides and homicides committed during armed robberies. The second
source, from police records, includes only criminal homicides. Since criminal homicides
are more frequent, especially in peak crime years, this makes only a modest difference as
can be seen for the two years when the two series overlap.
5

The chart shows a steady increase from 1980 to 2000, then a turning point and a
sharp decline. In 2001, there were 5162 criminal homicides in the city of São Paulo. In
2007, there were only 1527, despite an approximately 400,000 increase in the city’s
population. Monthly homicide statistics from the Secretaria de Estado de Segurança
Pública (unpublished) for the city of São Paulo show a steady decline for the entire
period from January 2001 through June 2008. There were 477 criminal homicides in
January of 2001 and only 97 in January of 2008. January is mid-summer in São Paulo,
a time when homicide and other crime rates tend to peak.
Most Brazilian statistical data are reported by state, and these data provide more
details on homicide trends. Chart Two compares trends in criminal homicide, attempted
homicide and negligent homicide in the state of São Paulo. The sharpest decline is in
criminal homicide. The rate for negligent homicide (96% of which are automobile
accidents) did not show a parallel decline; the decline was in willful, intentional murder
and attempted murder.
Comparative data for Brazil as a whole and for other jurisdictions in Brazil are of
uneven quality and not always as up-to-date as the São Paulo data. But the available data
do not show a Brazil-wide crime decline, at least before 2005. Data from the national
criminal justice statistical service, in Chart Three, show a stable criminal homicide rate
for Brazil as a whole from 2001 to 2005 (SENASP: 2005, 2006). During this period, the
rate declined only slightly in Rio de Janeiro, Brazil’s other huge urban agglomeration.
The fact that São Paulo trends differ from those elsewhere in Brazil suggests that policies
implemented on the state level are likely to have been responsible.
6

At 14 per 100,000 in 2007, the criminal homicide rate in São Paulo has not yet
quite declined to the levels achieved by New York City (7 in 2004). But it compares very
favorably with the rates reported by Detroit (42), Baltimore (44) and Washington, D.C.
(36) in the same year (Statistical Abstract, 2007). The American statistics are for
“murder and non-negligent manslaughter,” a category that is approximately comparable
to “criminal homicide” (homicídio doloso) in Brazil.
Disaggregating São Paulo Homicide Patterns. Using data for the city of São
Paulo in 2001, Gawryszewski, Kahn and Jorge (2004), found a remarkably strong gender
difference, with a homicide rate of 111.1 per 100,000 for men and only 7.4 for women.
The highest death rate was for 19 year old men. Fifty-six percent of homicide victims
were between 15 and 29 years of age. They found that firearms accounted for 66.5% of
the deaths as recorded on death certificates, and 88.6% of deaths recorded on autopsy
reports which are considered more reliable. The average number of bullet wounds per
victim was 6.9, with the largest number of wounds in the head. A more recent study by
the Secretaria de Estado de Segurança Pública (unpublished), using a sample from 2005
data, found the average number of gunshot wounds to be 4.5. Most of the victims in the
2001 study, 66.0%, were transported to a hospital, although not all received medical
assistance. Among those victims whose blood was tested, a little more than half, 55.8%,
were negative for alcohol. Very few, less than 1%, tested positive for cocaine. These
figures are likely to be high, since the police request blood tests when they have reason to
suspect alcohol or drugs are involved. They also found that the largest number of
homicides occurs on the weekend, with the peak on Saturday and the lowest number on
Wednesday.
7

Kahn (2004) observed an annual cyclical pattern in homicides and attempted
homicides in the state of São Paulo, peaking in the first trimester of the year which is
summer in the southern hemisphere. Ceccato (2005) found that most homicides in the
city of São Paulo took place on the evenings and weekends during the hotter part of the
year and in low income neighborhoods. She observed that the significant reduction in
homicides in recent years was not always accompanied by a similar reduction in other
violent crime, and suggested that effective gun control measures were a significant causal
factor.
Examining trends from 2000 to 2003, Kahn (2004) found that most of the
reduction in homicides was in the larger cities of the state, including the capital itself,
where the homicide rates are highest. Within the capital city, the homicide decline could
be observed in 75 of the 93 neighborhoods; it was not concentrated in any geographic
region. The reduction was observed both in homicides in public places and in homicides
within homes and commercial establishments.
Policing and Homicide Rates. The impact of policing efforts on crime rates
remains controversial and difficult to measure. In the 1970s, the consensus among
criminal justice researchers in the United States was that “nothing works” (Martinson,
1974; Kelling, et al, 1974). This remained the consensus opinion until the 1990s when
trends shifted and suddenly researchers began to find that “everything works” (Zimring
2007: 25-42). In periods of declining crime rates the pendulum may swing from
excessive pessimism to excessive optimism.
A more encouraging reading of the United States experience is that the police,
especially in New York City, learned from the failure of routine patrolling and other
8

traditional activities and adopted more effective tactics in the 1990s. In the most
thorough evaluation of the New York City experience, Zimring (2007: 151) concludes
that: “there is powerful circumstantial evidence that compound major changes in the
quantity of police and the tactics of policing had a major impact on crime.”
The police in São Paulo responded to the rising crime rates in the 1990s with
changes in management culture similar to those instituted in New York City in the 1990s.
They gave new priority to gathering accurate and timely empirical data and using it to
plan and evaluate programs. An intergovernmental communications network was
established to link the military and civil police. Crimes were entered into a geographic
information system, and saturation units were sent to areas controlled by drug traffickers.
A data base was established with photographs of over 300,000 criminals. Telephone
switchboards were set up to receive citizen complaints of incidents, and a web site was
opened to take reports of thefts of vehicles, documents and cellular telephones.
Community policing stations were opened, and a homicide combat unit was organized
with an emphasis on solving difficult cases. A specialized unit was organized to provide
supportive assistance to women who were victims of sexual crimes. Sophisticated
computer software linked information from police reports with bank records, telephone
records and probable areas of residence. And the police began more aggressive efforts to
remove illegal firearms from the streets.
As a consequence of these efforts, the number of imprisonments in the state of
São Paulo increased from 18,602 in the first quarter of 1996 to 30,831 in the first quarter
of 2001, after which it settled back to approximately 23,000 a month, as shown in Chart
9

Four. The turning point in the state’s criminal homicide rate came at the peak of this
increase in imprisonments, as shown in Chart Five.
Data on other crimes are not as good as the data on homicides, making it difficult
to determine whether the improvements in policing caused a generalized decline in
crime. Reported kidnappings shot up from 12 in 1996 to 307 in 2001, settling back to
123 in 2006. The very low number of reported kidnappings in 1996 suggests that most
victims chose not to report the crime to the police, since kidnapping was widely
believed to be high during this period. The dramatic increase in reports may reflect
confidence in the ability of special anti-kidnapping units to solve kidnappings. There
was a spurt in reports of carjacking in the second two quarters of 2002 which may
reflect increased belief in the ability of the police to recover stolen automobiles. The
reports of carjacking declined after that.
Arrests for drug use and drug trafficking have grown steadily in São Paulo since
2001. These arrests have contributed to prison overcrowding, increasing the pool of
young men available for recruitment into organized crime. In future, São Paulo
authorities hope to rely more on treatment than on imprisonment for drug offenses.
Poverty and inequality may play a larger role in theft and robbery than they do in
homicide, and these crimes have not declined in step with the decline in homicide. Most
criminal homicides involve conflicts between young men, often under the influence of
alcohol or drugs. Several São Paulo municipalities have closed bars early as a measure to
reduce violent conflict. Removing guns from the population seems to be highly effective.
Gun Control and the Homicide Drop. According to data from the Ministry of
Health, the principal cause of death from external causes in Brazil is “aggressions” for
10

men and transportation accidents for women. Death from aggressions is 12 times more
frequent for men than for women. Among men, 72% of aggressive deaths are caused
by firearms, as compared to 54% for women. In 2003, 33,991 Brazilian men and 3,937
women were killed by firearms (Saúde Brasil, 2005). In 2002, there were 21.7
firearms deaths for each 100,000 people, as compared to 10.7 in the United States
(Souza, et al., 2007: 575).
In October, 2003, the Brazilian federal government enacted a new set of laws to
limit the importation of firearms, make it illegal to own unregistered guns or to carry
guns on the street, and increasing the penalties for violation of gun control laws. In
2005, Brazil’s leading political parties and advocacy groups promoted a national
referendum to ban commerce in arms and ammunition altogether. Despite support from
all sides of the political spectrum, this referendum was defeated by a hastily organized
pro-gun coalition that argued that gun control would only deny guns to law abiding
citizens, making them sitting ducks for criminals.
Despite the failure of the referendum, Brazilian gun control legislation is strong
and some analysts (Souza, et al, 2007) attribute the recent drop in homicide deaths to
the 2003 legislation. Data from the Ministry of Health shows that “firearms deaths” in
Brazil increased steadily from 1992 to 2003, then turned down significantly (Painel de
Indicadores, 2006, p. 42). The Ministry of Health data include all categories of gun
deaths, including accidental deaths.
Unfortunately, there is a problem with the health department data on firearms
deaths, at least for some years in the state of São Paulo. Many physicians apparently
recorded firearms deaths as caused by an objeto contundente, an unfortunate translation
11

of the World Health Organization category of death by a “blunt object.” A bullet can
be thought of as an objeto contundente in Portuguese but not as a “blunt object” in
English. This confusion was apparently cleared up after 1999, at which time the
number homicides recorded as caused by an objeto contundente declined very sharply.
Because of this problem, we have combined firearms deaths and “blunt object” deaths
in Chart Six.
The most consistent data in Chart Six are those for deaths by knives and sharp
objects. Deaths by gunshot and “blunt objects” (many of which were bullets) peaked in
2001 and then declined significantly. The unfortunately high percentage of “not
specified” deaths further clouds the value of this data series. The fact that deaths by
knives and sharp objects are constant while other deaths have declined gives some
support to the thesis that firearm control has been an important factor.
The data from the federal Ministry of Health also show great variation from state
to state which also may reflect problems in reporting. Comparing 2003 to 2004,
reported firearms deaths declined 19% in the state of São Paulo, 9.9% in the state of
Rio de Janeiro, 14.5% in Pernambuco, and 20.6% in Matto Grosso. But they increased
by 7.2% in Minas Gerais, 29.3% in Amazonas, and 11.4% in Pará (Evolução da
Mortalidade, 2007). The reported improvement in the statistics for Brazil as a whole
can be largely accounted for by a very sharp drop in São Paulo which accounts for
about 25% of the reported national firearm deaths. The most important factor does not
seem to be the passing of national legislation, but the vigor with which the legislation is
enforced on the state level.
12

In the state of São Paulo, firearms confiscations by the police rose from 6,539 in
the first quarter of 1996 to 11,670 in the second quarter of 1999. This peak coincides
with the beginning of the great São Paulo homicide drop. Firearms confiscations
remained high through 2004, and then settled back to their previous level (Chart
Seven). São Paulo authorities believe that the decline in firearms confiscations after
2004 was because the new national legislation had increased the penalties for carrying
firearms and fewer persons risked carrying them on the street.
The São Paulo Homicide Drop in Comparative Perspective. The dramatic
drop in the São Paulo murder rate conflicts with media imagery. News coverage has
dramatized brazen attacks by organized criminals on police stations and public
transportation in the city of São Paulo as well as in Rio de Janeiro and other Brazilian
cities. These attacks paralyze a city for a day or two and contribute to Brazil’s image as a
dangerous place that many fear to visit. They are intended to generate media coverage,
embarrass officials and intimidate law enforcement. They disrupt the life of the
community and threaten the forces of law and order. But the number of people killed in
these attacks is small compared to the monthly toll of mundane homicides that do not
receive so much media attention.
The attacks by organized crime are a response to police crackdowns that have put
large numbers of offenders in crowded prisons and removed thousands of handguns from
circulation. These police measures have substantially lowered homicide and some other
violent crime rates. This does not generate dramatic news stories or television footage,
but it does make life much safer for the average citizen or visitor to São Paulo.
13

Most research on crime drops has focused on the United States (Blumstein and
Wallman 2006; Zimring 2007), examining trends over time and differences between
states and cities. Despite the plethora of research, the causal factors remain controversial.
In the most comprehensive review of this research, Zimring (2007) argued that data from
one country do not provide sufficient variation in many variables to adequately test
causal models. He added Canada as a comparison case, and found that Canadian crime
trends were remarkably similar to those in the United States although the trends in many
of the hypothesized causal factors were quite different.
Zimring and others have found that the 1990s crime decline was much sharper in
the city of New York than elsewhere in the United States. This fact is often attributed to
effective policing measures, including anti-gun measures, instituted by the New York
City police (Karmen, 2000). In the Brazilian case, the most relevant differences in crime
policies are between states, and these have made a tremendous difference in homicide
rates. The modernization of the police forces in the state of São Paulo has been highly
effective, and has had a major impact on the national statistics, just as the success in New
York City has been a major contributor to national statistics in the United States.
The Brazilian constitution has no provision guaranteeing the right to bear arms,
and Brazil was able to implement strong national gun control legislation in 2003,
something which has not been possible in the United States. The results suggest that the
legislation was helpful in reducing homicide, but only when actively enforced by the
police forces in a state. This is not primarily because professional criminals and drug
gangs were disarmed; it is largely due to a sharp decline in homicides resulting from
altercations between young male acquaintances.
14

One controversial hypothesis in the United States has no application to Brazil.
There was no change in abortion policies or practices in Brazil a generation before the
São Paulo homicide decline, so there is no possibility that the Brazilian crime drop was
due to an increase in abortion. This, combined with Zimring’s finding of a lack of age-
related homicide declines correlated with abortion legalization in several European
countries, suggests that any correlation between changes in abortion policies and
declining homicide rates a generation later in the United States may have been a
coincidence.
Drug use, on the other hand, is a problem in Brazil as it is in the United States. It
may be that the homicide explosion came later in Brazil than it did in the United States
because the crack cocaine epidemic came later. Confiscations of cocaine and crack have
continued to increase in São Paulo over the last five years, but it is not certain whether
this is due to more effective policing or to an increase in the amounts being sold.
Zimring (2007: 197) concludes his review of the crime decline in the United
States with the statement that “the crime decline of the 1990s was a classic example of
multiple causation, with none of the many contributing causes playing a dominant role.”
Similarly, Morrison and Bronkhorst (2006: 9) conclude that “there is no single solution
to reduce levels of crime and violence in Brazil.” Success has many fathers, and there is
enough good news in the São Paulo homicide decline to credit many of them. Most
importantly, the great São Paulo homicide drop shows that effective measures can be
taken to reduce lethal crime without waiting to solve underlying socioeconomic
problems.
15

Chart One
Homicides Rates per 100,000 in the City of São Paulo
0
10
20
30
40
50
60
70
80
1975 1980 1985 1990 1995 2000 2005 2010
Sources: Federal Ministry of Justice and São Paulo Secretariat of Public Security
A
ll
Homicides
Criminal
Homicides
16

Chart Two
Homicide in the State of São Paulo: 1996 to 2006
0
5
10
15
20
25
30
35
40
1994 1996 1998 2000 2002 2004 2006 2008
Source: Secretaria de Estado de Segurança Pública
Rate per 100,000 inhabitants
Criminal Homicide
A
ttempted Homicide
Negligent Homicide
17

Chart Three
Criminal Homicide Rates in Brazil, São Paulo and
Rio de Janeiro: 2001-2005
0
10
20
30
40
50
60
2000 2001 2002 2003 2004 2005 2006
Source: http://www.mj.gov.br/senasp/estatisticas
Rates per 100,000
Brazil City of São Paulo City of Rio de Janeiro
City of São Paulo
City of Rio de Janeiro
A
ll of Brazil
18

Chart Four
Quarterly Imprisonments in the State of São Paulo
0
5000
10000
15000
20000
25000
30000
35000
Jan-93 Oct-95 Jul-98 Apr-01 Jan-04 Oct-06 Jul-09
Source: Secretaria de Estado de Segurança Pública
19

Chart Five
Prisoners and Criminal Homicide in the State of
São Paulo
0
20
40
60
80
100
120
140
160
180
1994 1996 1998 2000 2002 2004 2006 2008
Source: Secretaria de Segurança Pública
2000 = 100
Prisoners
Criminal
Homicides
20

Chart Six
Firearm, Blunt Object, Knife and Unspecified
Homicides in the State of São Paulo
0
2000
4000
6000
8000
10000
12000
14000
1994 1996 1998 2000 2002 2004 2006
Source: Unpublished Health Department Data
Firearm and "Blunt Object" Knife or Sharp Object Not Specified
21

Chart Seven
Quarterly Firearms Confiscated in the state of São
Paulo
0
2000
4000
6000
8000
10000
12000
14000
Jan-93 Oct-95 Jul-98 Apr-01 Jan-04 Oct-06 Jul-09
Source: Secretaria de Estado de Segurança Pública
22

References
Andrade, M.V. and M. B. Lisboa. 2000. “Desesperança de Vida: Homicídios em Minas
Gerais, Rio de Janeiro e São Paulo no Período 1981/97. In Ricardo Henriques, ed.,
Desigualdade e Pobreza no Brasil. Rio de Janeiro: IPEA.
Baierl LF. 2004. Medo Social: da violência visível ao invisível da violência. São Paulo:
Cortez.
Beato Filho, CdC. 1999. “Políticas públicas de segurança e a questão policial.” São
Paulo em Perspectiva. 13(4):. 13-27.
Blumstein, Alfred. 2006. “Disaggregating Violence Trends,” Pp. 45-96 in Blumstein
and Wallman, 2006.
Blumstein, Alfred and Joel Wallman. 2006. The Crime Drop in America, Revised
Edition. New York: Cambridge University Press.
Brinkhoff, T. 2006. The Principal Agglomerations of the World.
http://www.citypopulation.de/World.html.
Casa de Nariño. 2006. “Cede Homicídio em Colombia,” March 12.
http://www.presidencia.gov.co/sne/2004/abril/06/17062004.htm.
Cavalcanti, Roberto and Sonia Rocha. 2006. , “Vez e Voz aos Pobres: Uma Nova
Politica Social,” paper parepared for the XIII Forum Nacional, May 2006.
www.inae.org.br/publi/ep/EP0139.pdf.
Ceccato, Vania. 2005. “Homicide in São Paulo, Brazil: Space-Temporal and Weather
Variations,” Estudos Criminológicos 3: 11-30.
http://www.ssp.sp.gov.br/estatisticas/downloads/manual_estudos_criminologicos_3.pdf.
Cerqueira, Daniel and Waldir Lobão. 2003a. “Determinantes da Criminalidade: Uma
Resenha dos Modelos Teóricos e Resultados Empíricos.” Rio de Janeiro: IPEA.
http://www.ipea.gov.br/pub/td/2003/td_0956.pdf.
Cerqueira, Daniel and Waldir Lobão. 2003b. “Condicionantes Sociais, Poder de Polícia
e o Setor de Produção Criminal, Rio de Janeiro: IPEA.
http://www.ipea.gov.br/pub/td/2003/td_0957.pdf
.
Coelho, Edmundo. 1988. “A Criminalidade Urbana Violenta,” Dados: Revista de
Ciências Sociais 31(2): 145-183.
DiIulio, John. 1996. Body Count. New York: Simon and Shuster, 1996.
23

Drummond Júnior, Marcos. 2002. Vida e Morte em São Paulo. São Paulo: Brasilense.
Evolução da Mortalidade. 2007. Ministério de Saude, “Evolução da Mortalidade por
Violência no Brasil e Regiões,”
http://portal.saude.gov.br/portal/svs/visualizar_texto.cfm?idtxt=24448.
Gawryszewski, Vilma Pinheiro, Túlio Kahn and Maria Helena Prado de Mello Jorge.
2004. Homicídios no Município de São Paulo. Estudos Criminológicos 1: 4-12.
http://www.ssp.sp.gov.br/estatisticas/downloads/manual_estudos_criminologicos_1.pdf.
Geração Online. 2004. “Entrevista com Luis Mir.”
http://geracaobooks.locaweb.com.br/releases/?id=29.
Hinton, Mercedes. 2006. The State on the Streets: Police and Politics in Argentina and
Brazil. Boulder: Lynne Rienner.
IBGE. 2004. Institute Brasileiro de Geografia e Estatística. “Projeção da População do
Brasil por Sexo e Idade para o Período 1980 - 2050. Revisão 2004.
http://www.ibge.gov.br.
IBGE. 2007. Institute Brasileiro de Geografia e Estatística. “Estimativas das
Populações Residentes em 01.07.2006, Segundo os Municípios. http://www.ibge.gov.br.
ILSD. 2007. Instituto Latinoamericano de Seguridad y Democracia. “Homicidios.”
http://www.ilsed.org.
Kahn, Túlio. 2004. “Homicídos Dolosos em São Paulo.” Estudos Criminológicos 1:
15-32.
http://www.ssp.sp.gov.br/estatisticas/downloads/manual_estudos_criminologicos_1.pdf.
Kahn, Túlio. 2004b. “Segurança Publica e Trabalho Policial no Brasil,” Estudos
Criminológicos 2: 75-86,.
http://www.ssp.sp.gov.br/estatisticas/downloads/manual_estudos_criminologicos_1.pdf
.
Karmen, Andrew. 2000. New York Murder Mystern: The True Story Behind the Crime
Crash of the 1990s. New York: New York University Press.
Kelling, George, et al. 1974. The Kansas City Preventinve Patrol Experiment.
Washington, DC: Police Foundation. www.policefoundation.org
Kilsztajn, Samuel, et al. 2005. “Taxa de Homicídio por Setor Censitário no Município
de São Paulo,” Estudos Criminológicos 3: 1-10.
http://www.ssp.sp.gov.br/estatisticas/downloads/manual_estudos_criminologicos_3.pdf.
24

Marinha de Souza, et al. 2007. “Reductions in Firearm-Related Mortality and
Hospitalizations in Brazil after Gun Control,” Health Affairs 26: 575-584.
Martinson, Robert. 1974. “What Works?” The Public Interest, Spring 22-54.
Mendonça, Mario. et al. 2003. “Criminalidade e Desigualdade Social no Brasil,” Rio de
Janeiro: IPEA. Texto para discussão número 967.
Mir, Luis. 2004. Guerra Civil: Estado e Trauma. São Paulo: Geração Editora, 2004.
Olsén, Örjan, et al. 2004. “Desemprego, Rendimentos e Crime: Um Estudo no
Município de São Paulo. Estudos Criminológicos 2: 4-73.
http://www.ssp.sp.gov.br/estatisticas/downloads/manual_estudos_criminologicos_2.pdf.
Painel de Indicadores. 2006. “Determinantes Sociais de Saude,” Pp. 36-47 in Ministério
de Saude, Painel de Indicadores do SUS
http://portal.saude.gov.br/portal/arquivos/pdf/painel_%20indicadores_do_SUS.pdf .
Paixão, L.A. “Crime, Controle Social e Consolidação da Democracia: as Metáforas da
Cidadania.” In: Reis, F. W. e O'Donnell, G. (Org.). Pp 168-189 in F.W. Reis and G.
O’Donnell, editors, A Democracia no Brasil. São Paulo: Vértice.
Pezzin, Liliana. 1986. Criminalidade Urbana e Crise Econômica: o Caso de São Paulo.
São Paulo: IPE/USP.
Romero, Simon. 2006. “As Crime Soars for Venezuela, Chávez Coasts.” New York
Times, December 2, 2006.
Sapori, Luis and C.B. Wanderley. 2001. “A relação entre desemprego e violência na
sociedade brasileira: entre o mito e a realidade”. Pp 42-73 in Violência do Cotidiano.
São Paulo, Fundação Konrad Adenauer.
Saúde Brasil. 2005. “Análise da Tendência de Morte Violente,” pp. 591-639 in Saúde,
Saúde Brasil. http://portal.saude.gov.br/portal/arquivos/pdf/saude_brasil_2005.pdf
.
SENASP. 2003. Secretaria Nacional de Segurança Pública. “Análise Comparativa do
Número Total de Vítimas de Homicídio entre as 26 Regiões Metropolitanas - 1980 a
2002.” http://www.mj.gov.br/senasp/estatisticas/homicidios/estat_homicidios1.htm
.
SENASP. 2004. Secretaria Nacional de Segurança Pública. “Mapa de Ocorrências no
Brasil 2001-2003: Brasil, Unidades da Federação e Regiões Geográficas.”
http://www.mj.gov.br/senasp/estatisticas/estat_ocorrencia.htm.
SENASP. 2006.. Secretaria Nacional de Segurança Pública. “Análise das Ocorrências
Registradas pelas Polícias Civis (Janeiro de 2004 a Dezembro de 2005).
http://www.mj.gov.br/senasp/estatisticas/mapacrime/Mapacrime2004_2005.pdf
.
25

SESP. 2007a. Secretaria de Estado da Segurança Pública, Governo do Estado de São
Paulo. “Estatisticas.” http://www.ssp.sp.gov.br/estatisticas/.
SESP. 2007b. “Evolução da Criminalidade 1996 a 2006 e Comparação com Demais
Estados.” Powerpoint Presentation.
SESP. 2008. Previously unpublished statistics from the data base of the Secretaria de
Estado da Segurança Pública, Governo do Estado de São Paulo.
Souza, Maria de, et al. 2007. “Reductions in Firearm-Related Mortality and
Hospitalizations in Brazil after Gun Control,” Health Affairs 26: 575-584.
Statistical Abstract. 2007. The Statistical Abstract of the United States.
http://www.census.gov/compendia/statab/law_enforcement_courts_prisons/crimes_and_c
rime_rates/.
Wilson, James Q. 1995. “Conclusing Essay in Crime,” Pp. 488-507 in James Q. Wilson
and Joan Petersilia, eds., Crime. San Francisco: ICS Press.
World Bank. 2006. Crime, Violence and Economic Development in Brazil.
Washington: World Bank Report no. 36525.
http://www.unodc.org/pdf/brazil/Crime_and_Violence_jan_2007.pdf.
Zimring, Franklin. 2007. The Great American Crime Decline. New York: Oxford.
26