Claudio Beato
Universidade Federal de Minas Gerais
Rising Violence in Brazil
The lack of solid empirical and factual base has compromised the theoretical discussions about the rising crime rate in Brazil and Latin America. The government agencies in charge of producing statistics about violence and crime have not invested in the creation of national database that can reliably and validly account for and describe violent events. This has made it difficult to obtain information about offenders, incidents, prevalence and victimization. As a result, most of what has been written about criminality and violence in Brazil are essays with little empirical support and short theoretical range. If we were to group the research on crime and violence in Brazil along a continuum that correlates the degree of technical elaboration of the analyses, the bulk of research would fall in the range of essays or at best bivariate tables. More sophisticated empirical research, such as multivariate analyses require data collection with a large number of cases in order to make accurate inferences. This data is not available in Brazil.
Other indicators, such as those in the fields of education, economics and health care are available. In the past twenty years, Brazil has made progress in such quality of life indicators as life expectancy, child mortality, literacy and income. Improvements have also been made in overall development, as shown by increases in the Human Development Index of the United Nations. Other improvements have been made on the political sphere, with free elections now in effect for fifteen years and political institutions that have survived various economic crises.
However, in this same period, there has been a significant increase in violent crimes, especially in urban centers. The homicide rate (per 100,000) has tripled in Rio de Janeiro since the 1970's, the rates in São Paulo and Porto Alegre have quadrupled in this time and Belo Horizonte has seen a rise of 50% a year in violent crimes, especially in recent years. The largest crime growth has been in violent crimes against property.
The rise in crime has significantly impacted the Brazilian economy in public security expenses, that in Rio de Janeiro are estimated at $1.7 billion a year, the equivalent of 5% of the city's income. This expense includes medical care for victims, premature retirement, security systems, justice and insurance fees.
Some Myths About the Causes of Violence in Brazil
Some would argue that the crime problem would be solved if we were to solve our economic and social problems. I have some good evidence that would not happen. A number of indicators have said to be related to crime, such as: poverty, the socioeconomic structure, and unemployment.
Looking at the distribution of poverty and crime, there is no significant correlation between violent crime and measures of either relative or absolute poverty, as measured by the Gini index or those below the poverty line. The same lack of relationship can be seen when looking at violent crime rates and the amount of public services. There is no clear relationship between the number of official institutions per 100,000 and the rate of violent crime. Examining unemployment and crime over time in the city of Belo Horizonte, no significant correlation can be found between the unemployment rate and the violent crime rate.
Crime and Development
While no significant correlation could be found between indicators such as poverty and unemployment and crime, there is a strong relationship between other indicators of modernization and development and crime. One such relationship is city size. In the state of Minas Gerais, the larger the city, the larger the crime rate.
The Brazilian development process is not different from what has happened in other countries. Urbanization and growth of big cities has always been related to growth in crime rates (see Glaeser and Sacerdote, 1995). This has been explained by the literature through the anonymity that the city provides, as well as a lessened degree of control over people (Wilson and Herrenstein, 1993).
More than this, the hypothesis that associates urban lifestyle as a component of criminality has been getting stronger. This has to do with the use of drugs, among other things. In fact, when we associate violent crime occurrences, particularly against property, we see that there is a very robust correlation of .55 between the use of drugs detected by the police and the incidence of crimes against property.
The point I am developing is that modernization and economical social development in Brazil will not necessarily solve our problems of crime and violence, but, in fact, probably will make them worse. When we look at Brazilian social indicators, we see how the last decades translate in terms of crime. (see Table 1).
Table 1
Correlation of Violent Crimes and Socio-economic Indicators
Socio-economic Indicator |
Violent Crimes |
Intra-personal Crimes |
Property Crimes |
| Children out of school | -.08 |
.13 |
-.35 |
| Index of human development | .06 |
-.22 |
.40 |
| Income index | .12 |
-.16 |
.39 |
| Education index | .03 |
-.22 |
.38 |
| Average years of study | .15 |
-.17 |
.47 |
| People working in transformation industry production | .38 |
.12 |
.57 |
| People working in Services | .36 |
.13 |
.53 |
| Gross Product Industry | .37 |
.14 |
.53 |
| Gross Product in Services | .28 |
.10 |
.43 |
| Gross Product per capita | .17 |
.02 |
.27 |
However, this correlation between social and economic indicators have shown only a part of the problem. The other has to do with the financial crises and the need for reforms in many different sectors, including public security. In Minas Gerais, for example, there were 12 arrest made for every 100 reported armed robberies. This number decreased to 8 arrests in 1994, and to 6 for 1997. This shows a drop in effectiveness in the police force that is related to poor preparation, poor equipment and poor salaries. Fifty percent of police officers are under 30 years old. Police are given only a short 8 month training period before beginning street patrol. Police firearms are 25 years old and patrol cars are 8 years old, on average. At any given time, 25% of patrol cars are broken.
"Opportunity Context", Motivations and Crime
I would like to build upon the idea that social and economic development could lead to a context of more opportunities for incidences of crime. As I have said, this is not a new idea if we look at the classical literature on this theme. But certainly it is new if we consider explanation in Latin American and Brazilian literature for the growth of crime rates.
Actually I don't believe that it is possible to find a single explanation for the rising rates of crime, because we haven't one single cause, but multiple causes, either in combination or interaction. But it is possible to go beyond some taken for granted explanations when we have good data.
In the Brazilian case, I think that it is time that we look more and more to some recent developments in criminological theory, especially those related to rational approaches like Rational Choice or Ecology theories. My data suggest that it is possible to build a model with macro-level variables to explain crime rates in a context of socioeconomic development.
The first step to be taken in this direction is to separate the "crime incidence" form "social characteristics of the criminal". Studies about Brazilian violence usually focus on the criminal'' social characteristics as a good predictor for the violence committed.
In theoretical terms, the theories that deal with opportunities state that 'delinquents are not different from non-delinquents': both are equally predisposed to crime.
Actually, the motivation can be seen as a result of stimuli specific of a given environment of action and oriented toward specific types of criminal acts. The literature on "Situational Crime Prevention" sets a clear distinction between motives (i.e. oriented to specific targets) and motivations (i.e. referring to generalized needs or drives, with specific targets).
This specific action environment, nevertheless, has to do with a larger socioeconomic macro-structure context, or, to put it into terms of the Routine Activity Approach, with the availability of targets for criminal action as well as the lack of guardians. On the micro-level we jump to the conclusion of the immediate environment of action as a context of deliberation. On the macro-level, this situational availability relates itself to the development of a socioeconomic structure, as I have sought to demonstrate here through data from Brazil.
Opportunity and motivations are the key elements to understand the process of rising crime rates in a context of social development. Opportunity is related with ecological variables. Motivations have to do with the social environment where the preference structure is formed. More than simple explanations linking poverty or development with crime, explaining crime through opportunity and motivations may provide more answers to the rising crime rate in Brazil.
Roberto Briceño-León
Universidad Central de Venezuela
The Relevance of Violence in Latin America
In almost all of Latin America violence is among the five top causes of death, and is the first cause of death in several countries: Brazil, Colombia, Venezuela, Salvador, Mexico. In Colombia, for example, between 1986 and 1996, violence caused more deaths than did HIV in the whole of Latin-America.
Another way of measuring this rising tide of violence is the years of life lost calculation. For example in Mexico City (the DF) 57,673 years of healthy life were lost in 1995. Of these 68.4% were due to premature deaths and 31.6% due to injuries; while in Sao Paulo homicides were responsible for 3,483 years of life lost for each one million of inhabitants according to one source. In Rio de Janeiro the potential years of life lost by violence represents the 19% of the total years lost, but cancer represents 10.3% and cardiac diseases 7.2%. However, morbidity data relating to injuries from violence are rarely accurate. Nevertheless, according to the Pan-American Health Organization 30% of the hospital admissions in the region are tied to violence.
Some Characteristics of the New Violence in Latin America
First, with one or two exceptions (e.g. Colombia) it is generally a new phenomenon. The increase in the rate of homicides in Brazil and Venezuela started during the eighties, and in Mexico at the beginning of the nineties. In Venezuela the homicide rate was 9/100000 during the sixties, seventies and the eighties, but increased to more that 20/100000 at the beginning of the nineties. In Mexico City in 1987 homicides represented 1.4% of the total deaths, in 1991 2.8%, and in 1995 was 3.3%.
Second, it is largely urban. Despite the preponderance of rural violence in the past, today it is higher in big cities, such that Sao Paulo and Rio de Janeiro, Mexico City, Lima and Caracas concentrate more than half of the total homicides of the country.
Third, the new wave of violence is not political, but tends to be related to other "ordinary" crime such as robbery. In Chile between 1983 and 1992 the robbery with violence increased from 30% to 56% of total crimes, while in Mexico City in 1991 homicides linked with theft were 38% in men and 26% in women, and increased in 1995 to 47% in men and 32% in women. In Venezuela while the theft of cars without violence has decreased in the past 10 years, theft with violence has increased.
Victim Profiles:
Most victims are male -- 90% generally. Moreover, most are relatively young -- in the 15-35 age range; and are poor. Although the middle class feel that they tend to be the primary victims, in reality the actual victims of violence are poor people living in shantytowns. For example, in Rio de Janeiro the homicide rate of people between 15 and 34 years of age in low income areas as Rio Comprido and in Sao Cristovao was 144/100000 and 177/100000 respectively; while in middle or high income inhabitants areas it was much lower (Copacabana, 56/100000; Botafogo, 28/100000; Lagoa, 23/100000. In Caracas more than 85% of the now all too frequent weekend homicides occur in the shantytowns areas (barrios).
Fear of Violence is Widespread and has Major Repercussions:
Today, the fear of violence is high within the population, and is particularly focused upon certain areas of the city and on public transportation. This has a number of important consequences: 1) inhibition and the sense of "losing" access and rights to the city; 2) so called "hard-targeting" and rising expenditure of public and private money in security; 3) a growing disposition to carry guns; 4) growing support for the death penalty; 5) growing support for violent and sometimes illegal behavior of the police; 6) support the right to kill -- particularly in self-defense, in defense of one's family, in defense of one's property, and even the right to kill the abuser or undesirable person.
The Study about the Right to Kill
The study was carried out in eight cities and was applied randomly to a sample of respondents between 18 and 70 years of age who were interviewed in their homes. A total of almost 11,000,821 valid interviews were collectedwith modest variation in totals between the cities of Salvador(Bahia Brazil), Rio de Janeiro (Brazil), Santiago (Chile), Cali (Colombia), San José (Costa Rica), San Salvador (El Salvador), Madrid (Spain), and Caracas (Venezuela). The rates of non-response varied between 6% and 35%., and the confidence level of the survey was 95%.
The questionnaires sought to elicit agreement or disagreement with a number of statements (Likert scales) relating to the right to kill, and the statistical analysis involved cross-tabulating two variables and calculating odds and odds ratios for different categories of variables of interest. An odd is defined as a quotient of probabilities, and in particular the probability of an event A in relation to the probability of a complementary event Ac; it is estimated through the corresponding quotient of relative frequencies.
The results suggested important differences by:
Looking across the cities the results suggest that the highest levels of approval for the right to kill were in Caracas and the lowest in Madrid, and generally the greatest discrepancies over the justification for the right to kill is for "social discrimination" killings, followed by killing of individuals who threaten the community. The converse is that there is the greater consensus (i.e., the greatest similarity among all the cities) for killing in defense of family, followed by killing a daughter's rapist.
Conjectures About the New Wave of Latin American Violence
Three different kind of factors that may help us to better understand violence are:
Emilio Dellasoppa
University of the State of Rio de Janeiro
Alicia M. Bercovich
IBGE-Brazilian Institute for Geography and Statistics
Eduardo Arriaga
Georgetown University
Introduction
What factors could help to explain how the homicide rate in the Municipality of São Paulo increased from a figure of 5 homicides for every 100 thousand inhabitants to almost 50 homicides per 100 thousand inhabitants between 1960 and 1997? What influence does this have on some demographic indicators, such as life expectancy and the number of years lost?
The point is that the level of 5 homicides per 100 thousand inhabitants was already a historic one, and has been an indicator of the Brazilian demographic structure for at least two decades prior to 1960: in 1946, when G. Mortara analyzed the death rate due to external causes in1940 for the Municipalities of Rio de Janeiro and São Paulo, he found those very same levels: 5.4 per 100 thousand inhabitants in Rio de Janeiro and 4.0 for São Paulo. The current figures are ten to fifteen times greater. They must be related to the transformations in Brazilian society over the last decades.
In the Municipality of São Paulo, the homicide rate grew 158,6 %. It was 17.4 in 1980 and 45 per100 thousand in 1991. This increase, which is far more than that of the two previous decades (91.2 % and 75 %), reflects qualitatively different levels of violence than those that emerged in the 1980s. In the twenty-one years between the censuses of 1991 and 1970, the homicide rate in the Municipality of São Paulo increased 395 %. The explanation for these phenomena must be sought at a socioeconomic level, and in the transformations of the structure of social relations during the period. It is enough to observe corresponding figures on a national level to realize the gravity of the situation: There were 11.7 homicides for every 100 thousand inhabitants in Brazil in 1979, a figure that increased to 21.0 in 1994,4 and 27.0 in 1996. 5
The high inflation rate that characterized the 1980s (often referred to today as the "lost decade") up until the present day, rising unemployment rates, and social inequality are factors to be considered. All violence-related indicators, such as mortality due to external causes, violent crimes, and homicides, rose. The differential between gender also grew. For example, the male homicide rate in São Paulo City rose 173.1 % between 1980 and 1985, and a further 24.6 % between 1985 and 1990.
This work analyzes the evolution of years of life lost due to external causes according to age and gender between the ages of 15 and 64 in the State of São Paulo and the Metropolitan Region between 1980 and 1996. An analysis of characteristics, specific and decisive to the structure of Brazilian social relations, will be made.
Data and Methods
The data used on deaths due to external causes and homicides comes from SIM Information System on Mortality (Ministry of Health), Vital Statistics provide by the IBGE, and from Brazilian Censuses. The codes for cause-of-death used by SIM are Codes E in the Ninth Revision of the International Classification of Diseases (CID-BR). Averages to the order of three years were used so as to minimize statistical variation in the data.
To evaluate changes in deaths due to natural and violent causes, the changes in death rates by causes and age groups between 1981, 1991, and 1996 were analyzed. The method of years of life lost is that of Arriaga 6, to provide a quantitative evaluation of violent and non-violent deaths. This method allows one to analyze changes in mortality according to individual cause and age groups, as well as for the whole of combined age groups. The number of years lost doesn't directly measure the change in mortality, but rather the impact on a population's life by the change in mortality, as well as life expectancy at birth.
The method of years of life lost takes into account the age at which a person dies, and directly refers to the changes brought to life expectancy. It is the measure of the number of years of life, or of the temporary life expectancy between the ages analyzed that are lost due to violent or non-violent causes. The difference between the years of life lost in a same geographical area from one year to the next indicates the variation in the temporary life expectancy incurred by a change in mortality. If one were to consider different areas (and therefore different populations), the difference in the number of years of life lost between the two populations reveals the change in the temporary life expectancy which would have taken place had this particular population's level of mortality due to a specific cause remained at the same level as that of the other.
For example, when comparing the Metropolitan Regions of Rio de Janeiro and São Paulo, the difference in years of life lost would show the increment in the temporary life expectancy between the ages of 15 and 64 if the mortality of the MRRJ were equal to the mortality observed in the MRSP. An analysis from ages 15 to 64 was made and subsequently a comparison drawn with the present mortality supposing a death rate of zero between these ages. The study focuses on these ages because young men are the group most affected by violent deaths.
Results
Figure 1 shows the evolution of the number of years of life lost (YLL) due to all causes, violent causes and natural causes, for males and females and age groups in the Metropolitan Region of Rio de Janeiro, in 1980, 1991 and 1996. When considering the total causes of death, we can observe a stabilization in the total values between 1991 and 1996 for all age groups, at slightly lower values than 1991. As usual, the impact of YLL for violent causes is weak for females.
For males, as shown in the left side of Figure 1, the calculation of YLL for age groups in the Metropolitan Region of Rio de Janeiro, clearly confirms that death for violent causes is concentrated in males of age groups 15-19 and 20-24. It must be noted that for these two age groups the relation in the number of YLL for males/females is almost 10:1, but, even in the case of the males, we observe a strong increase between 1980-1991, followed by a plateau between 1991-1996. (Or even a small decrease, that is more noticeable in the interior of the State of Rio de Janeiro). In this case is clear the scissors effect observed between the two inverse tendencies of natural and violent causes as we move into the older age groups.
In the paper we calculated the differences in the number of YLL by causes and age groups in the two time intervals considered, 1980-1991 and 1991-1996, (not presented here). The results are interesting because we can "read" processes that influenced natural and violent causes of death. We observe that improvements in health conditions in the 1980-1991 period for the three first age groups were cancelled or strongly reduced in the next period. We can induce the effect of AIDS on the 25-34 age group, in the Metropolitan Region of Rio de Janeiro, producing a reversion in the next period. The cohort 45-54 appears as an negative example of the process of decreasing yields in the combat against natural causes of death. At the same time, we have a reversion in the behavior of the number of YLL associated with violent causes in the interior of the State. The differences between the Metropolitan Region and the rest of the State of Rio de Janeiro, confirms the AIDS epidemic as a phenomenon strongly related to metropolitan areas.
Taking into account that we analyzed two time intervals of different duration, were calculated the differences in the mean yearly number of YLL for causes and age groups for males and females. These results show the same reversion in the behavior in the number of YLL associated with natural causes, in the Metropolitan Region of Rio de Janeiro, and a parallel reversion process associated with both natural and violent causes in the interior of the State of Rio de Janeiro.
These results show that in the Metropolitan Region of Rio de Janeiro, in the period 1991-1996 there occurred a decrease in the number of YLL by violent causes even for males. These reversion does not exist in the case of females. But in this case we must note that we are analyzing small values, resulting from the fact noted above about the relation between the number of YLL between males and females that for the initial cohorts stands about 10:1.
In the case of the São Paulo Metropolitan Region, we can observe in Figure 2 the evolution of years of life lost due to all causes, violent causes and natural causes, for males and females in the Metropolitan Region of São Paulo (MRSP), in 1980, 1991, and 1996. In the case of the number of years lost due to natural causes, the figures are related to changes in the age pattern of male mortality. We can see that due to AIDS the number of years lost of life grew among men between ages 15 and 44 in 1991. By 1996 already, the data indicates a drop in the number of years lost among men between the ages of 15 and 34, a stabilization in the figures of 1991 for the 35 to 44 year age group, and another decrease in the group of 45 to 54 years. The result is a notable drop in the number of years lost in men belonging to the lowest age group: 15 to 24, where the value (0.27) is below that of 1980. This probably indicates a good assimilation of means of protection, resulting in an improvement of more than 50 % over the figures of 1991. There is also an improvement in the 25 to 44 year age group, though much smaller, around 16 %, and no changes for males aged 35 to 44. It should be pointed out that death rates (males and females) due to AIDS in the Municipality of São Paulo for 1991 are always higher than corresponding death rates due to homicides, reaching a peak of 66 per 100 thousand inhabitants in 1995 before declining. The death rate due to AIDS for males is above that of homicides between 1992 to 1995. 7
Considering the number of YLL due to external causes for men in age groups from 15 to 64 in the MRSP, we observe that every age group, with the exception of 55 to 64, show significant increases in the number of years lost both in 1991 and 1996. In 1991, the result for young men aged 15 to 24 was even more dramatic, with an increase of more than 100 % above the figures of 1980. The number of years lost due to external causes in 1991 rose from 2.48 to 3.27. This change is significant if when compared to the figures for South Region, in 1991 (1.48 years) or for Argentina over the same period (0.79 years). The results of 1996 for men also show an increase, where the wave of years lost due to violent causes manifests itself with yet greater intensity in the following age groups: It is the 25 to 44 age group that climbs from 0.71 to 1.09 years lost, an increment of almost 56 %. The fact that the interval was of five years in this case as opposed to 11 in the preceding must also be taken into account.
The impact of the rise in male mortality in the 15 to 34 age group is such that it could influence or even reduce the life expectancy of the male group in its totality. Moreover, if we take into account the fact that the Brazilian population is currently passing through a new "youth wave", this could result in an exceptionally high number of violent deaths in the next few years.
In the right side of Figure 2 we present the same corresponding results for females. In the case of natural causes, in complete contrast with the male situation, female statistics have shown a tendency to diminish the number of years lost between 1980 and 1991 and a period of stagnation between 1991 and 1996. (Due to the different sizes of the intervals data cannot be directly compared). For violent causes, we can see that the risk of death due to violent causes is far greater for men, as opposed to the prevailing low figures for females.
Implications of the research and observation
If in the past, public health policies in Brazil essentially dealt with infectious diseases which represented the group of diseases that contributed the most to the total death rate, the changes in the patterns of mortality demography show, at present, an increasing importance of what are known as social diseases, which include external causes such as accidents or violence. Violence constitutes a matter of public health due to its impact on the health and well being of young people. Demography, by researching the geographical disaggregation of a phenomenon that varies from one region to the next, could be one of the scientific approaches to collaborate to the understanding of the problem of violence in Brazil. Demography can also contribute to the debate on death rates and indexes of violence, and help to determine what measure could effectively curb violence. In short, demography and social sciences have a role in any attempt to resolve these problems in Brazil by providing the necessary information to any attempt at a theoretical approach.
Marcelo Bergman
University of Oregon 8
Crime has risen sharply in the city of Buenos Aires. Despite the lack of reliable statistics, and serious methodological questions related to data collection, it is nonetheless safe to assume that the total number of property and violent crimes increased significantly over the last decade.
Problems with the Data
Argentina does not have a methodologically sound system of data collection on crime. Statistics from the police department are not published regularly, and there is a substantial suspicion that this data is not collected uniformly nor under strict professional guidelines. There is no professional scrutiny over the process of survey and compiled statistics design, nor any quality control processes of data collection has been established. Therefore, the scattered information reported by the police and the interior ministry should not be seriously considered for any systematic analysis.
An additional source of data is the compiled court statistics. Both the Supreme Court and the Justice Department have annual reports of cases filed and adjudicated by the system. This data does not provide a significant breakdown and dis-aggregation of the crime categories, therefore, there are numerous limitations to the quality of research that can be conducted on the specific patterns of criminal behavior.
According to a Justice Department Report for 1996 the number of non-narcotic violent crimes filed in federal courts increased by 13% over the previous 12 months. It should be remembered that those were filed cases, which do not portray a reliable estimation of the actual "crime map". As shown by Lynch (1995), many unaccounted factors affect the variations in reporting trends, which they may not reflect actual changes in crime activity. Nonetheless, a one year 13% increase of filed violent crime cases is significant, and it suggests a high likelihood of growing criminal activity.
In 1998 more than 900,000 cases filed in criminal courts. This figure doubles in 1997 450,000 cases. Among the people arrested, 77% were first timers, and 75% under the age of 19. However, we cannot predict with absolute certainty whether we are witnessing a new massive way of offenders. This data may simply reflect the inability of the police to apprehend more "professional" offenders. Victimization surveys provide the most accurate estimates of criminal activity, but does not provide an answer as to who are the offenders. The frequency of offenses can be estimated mainly through self-reports, however, those studies have never been conducted in Argentina.
Self Reports
Unfortunately, no rigorous data collection on self-reports exists to allow an analysis on crime. Some surveys on drug consumption or other minor violations are sparingly conducted, however the samples are generally small, the survey methodologies are usually unreported, and their design many times is questionable. Moreover, since neither the federal nor the provincial government have a systematic data collection to measure and design crime and prevention policies, such self-report data collections gathered unsystematically, which does not allow for comparisons over time.
The research division of the Justice Department is currently exploring self-report surveys on three specific areas. 1) Alcohol and Drug consumption, 2) Access to firearms, 3) Inmate surveys (to determine frequencies of offenses). It should be emphasized that those are in preliminary stages, and funding has still not been secured.
Victimization
As known from the criminological literature, victimization surveys provide good estimations for some crimes, given the fact that the majority of them are never reported to the authorities. However, sex and some other violent crimes are consistently under-reported, and white-collar crimes go generally unnoticed.
Victimization surveys could be a legitimate source of data only if they are methodologically sound. Samples must be large, questionnaires must be repeatedly tested, and particularly, contacts with the responded households are repeated over many pre-established periods. For instance, the U.S. each household is being visited 7 times (some of the contacts are conducted by phone) and the first interview data is actually dismissed. The sample in this country is 42,000 households per year, covering more than 84,000 respondents.
In Argentina, many victimization surveys were conducted lately under very serious questionable survey designs (Fraga 1998, Roemer 1997). Nonetheless, the more rigorous survey conducted by the Justice Department provide a clear sign of the rise in crime in the greater Buenos Aires area. For instance the "robo con armas" (a category closely associated with the UCR's category of robbery) increased by 28%, over a period of 18 months. The equivalent to burglaries increased by 24%, and "hurtos" (closely associated to larceny and theft) increased bye 23%. A cautionary note is needed. The samples in these 2 surveys are very small (1200 in the first, 1250 in the second), and some questions of randomness could be posed. Also, the surveys are based on first time interviews asking respondents to relate to the previous six months. Therefore, reliability and consistency problems call for a cautionary approach to this data.
Overall, the number of crimes reported to the police was 35% in 1998. This rate is compatible with the U.S. (38%). This relatively high rate of crime reports in the metropolitan area of Buenos Aires, however, raises several question marks. This rate is suspicious since the legal institutions in this country are generally not trusted, these institutions do not enjoy a good rate of confidence among the population, and the victims' expectations for prompt resolutions and reparations of harms generated by the committed crimes are very low. Therefore, Argentina's compatible reporting rates with the USA suggests that some problems with the victimization survey may exist. Respondents may be actually providing inaccurate responses. Therefore, even though those are the best estimators of crime, they should be considered carefully.
Other surveys illustrate tendencies in the crime activity. For instance in a recent survey of 800 people (Ambito 22-7-98) 62% of the respondents said that either themselves of a member of its nuclear family has been stolen over the last year.
Finally, some surveys of the general population had graded crime along the top three major concerns of the greater Buenos Aires population. Until 1994, the population major concerns were related to the economy, corruption, social services and education. The fact that several survey crimes had climbed to the third major concern in 1996, to the second in 1997, and in one recent survey crime became the first major concern, it portraits a populations' legitimate concern of this growing social problem. It suggests that the survey findings are not only a population's paranoia, a media selling issue, or a moral panic, but a real reflection of the rise in crime.
In sum, the scattered and unsystematic data that has been collected lately show a very clear trend of a significant rise in crime. Now the question is why?
Explaining the Rise in Crime
The very first and professional answer to that question is that until good data is collected no firm conclusions can be drawn. Generally, the answers are in the realm of speculations and of untested hypotheses.
An array of probable causes could be listed. The economic transformation and the growth of the underclass, the expanding group of "marginados" out of the recent economic transformation, the increasing number and availability of guns, the growing consumption of alcohol and narcotics etc.
The most commonly advanced proposition is the correlation between crime and unemployment. Since crime is rising while unemployment and the underclass is growing as well, it must be concluded that there should be a strong association between these two social processes.
The number of unemployed in the Greater Buenos Aires area doubled from 1986 to 1996. However, since the two best predictors of crime are gender and age, for crime purposes, the correlation between crime and unemployment needs to be based on the study of the unemployment rate in the male 18-25 category may have some effects on the criminal activity.
Based on the INDEC data for the Greater Buenos Aires area, the unemployment rate of males 18-25 category from 1986 until 1996 had almost quadrupled. For every three young unemployed male in 1986 there were 15 unemployed males in 1996. In terms of male High School dropouts, even though the data is somewhat incomplete and does not include the Federal District (1/3 of the total population), the percentage of High School dropouts has tripled over the same years.
The current rate of unemployment of 1998 for the 15-24 age category is 25%. Among the lowest quintile is 44% (under the poverty line) there are 116.000 (85% males) and in the fourth quintile and additional 90,000 youngsters. In what is considered poor households, one of every four youngsters are in this situation. This "marginality" condition increased since 1992 by 62%.
However, it should be reminded that correlation does not necessarily mean causation. In other words, there is a need for an explanation on how one variable, the unemployment, the "marginality" affects the other, i.e., crime.
Unfortunately, the study of Criminology in Argentina does not have an empirical tradition. Crime has always been conceived as a law related deviance problem overwhelmingly taught at law schools. Such legal tradition has biased the research and did not allow for the development of a quantitative orientation as in many other countries.
As we know, crimes are more than law violations, and the delinquent is more than a law-abiding citizen. Criminology theories developed considerably since Beccaria and Durkheim and even since Merton and Surtherland, however, as mentioned, the lack of rigorous research on crime in Argentina denotes a clear bias towards crime theories that are not grounded on facts but on philosophical and legal theories. Not surprisingly, the state of crime data, as described above, is absolutely chaotic. Today politicians and pollsters try to respond to new society's demands without firm grounded understanding of the cases and scope of the phenomena.
So why is crime rising in Argentina? Let me also speculate by pointing to a somewhat different theoretical approach. Very schematically, the criminal offender and the delinquents are generally people with very low self-control. The individual search for immediate gratification is not being neutralized by the possible costs such pursuit of gratification may entail.
Crimes are acts of pursue of immediate gratification, but are also facts that stem from opportunities, i.e., the availability of a desired object, or the opportunity that rises when a potential offender is "at the right place and at the right time".
Therefore, crimes occur when an individual with very low self-control becomes aware of the opportunities that he or she may have for committing a particular crime. In other words, in order for crimes to occur there is a need on the one hand for an offender, and on the other hand for the opportunity that actually triggers the plausibility for such act.
Is such hypothesis advanced by control theory holds. the best mechanism to reduce crimes are on the one hand better lockers, (i.e. to strengthen all the instruments that reduce the opportunity factor), and on the other hand the enhancement of youngster's self-control is in the particular family.
A strong association exists between a neglected and/or victimized child and a future potential offender. This is so because self-control in these youngsters is weak. The traditional power of institutions of control such as the morals, religion, community and the family had eroded. Such major social and cultural transformations have generated the conditions in which many kids and young adults grow up with marked levels of low self-control. Therefore they generally disregard the social and legal costs of crime. Moreover, when detection and punishment institutions for crime prevention and contention perform very poorly, as they do in Argentina, they only fuel more disregard on the already perceived lower costs of crime among weak self-control youngsters.
Over the last decades Argentina had major social and cultural transformations. massive incorporation of women into the workforce. The segregation and erosion of community. Rising expectations of citizenship and complete inability to meet these subjective expectations. Single headed households increase. Changes in the role of church.
In short, it seems to me that a profound change in many social institutions during the last 20 years in Argentina are to be studied in order to comprehend why there is more crime today than a decade ago. The precise connection between "marginality", unemployment, drugs and crime needs to be searched on the basis of how these social transformations generated people with low self-control.
Rapporteur: Sandra Torres (ILAS)
Before measuring crime rates and violence, it is important to first define violence: does it include property crime, homicide, domestic violence, etc? In particular, are traffic homosides included? Due to their distinct dynamics, different crimes must be kept separate and approached with tailored strategies for each.
We are experiencing severe changes in socio-economics, demographics, and the relationship between democratization and increased crime rates.
We are faced with a real dilemma, because social problems alone cannot account for the sudden rise in crime rates we are witnessing. Although crime has been tied to Democratization, we are not in a comfortable position to state that Democratization leads to increased crime rates. The fact that two events occur together does not mean there is correlation between them. For example, democratization is not necessarily correlated with police efficiency.
The relationship between factors affecting crime and the levels of crime is not linear, but complex. Crime rates are not directly related to the economy, and therefore, we cannot expect to decrease the crime rate by "fixing" the economy.
Certain "hot spots" of high crime rates are generally in wealthier areas, such as big cities and commerce centers, as can be observed in the U.S. However, crime rates depend on the units being looked at: When states or municipalities are looked at as a whole, the average crime rate is seemingly low; whereas looking within a city exposes the high crime rate of poor areas and the lower rate of wealthier areas. It is important to continue going further in explaining complexities such as this.
Youth have the highest risk of dying, as compared to other age groups. There are growing social problems, not unique to Latin America. Today, children everywhere are beginning to view their future as a very dim one. Even in video games, children are exposed to violence. We need to address this issue in all nations.
Are social problems and the breakdown of moral structure at the root of rising crime rates? Or could it simply be more complete data collection and processing, made possible by more advanced information systems.
The factor of corruption within the law enforcement must be addressed and accounted for. Though we cannot control police officers, we must strive to reform them through democracy. In Brazil, where police officers admit to being part of the problem, very few are prosecuted. In fact, less than 1.7 percent of criminals in Brazil are convicted. Therefore, the judiciary system is what needs reforming, as well.
There needs to be a way to separate real crime control from the "dark side" of law enforcement, meaning crimes committed by police officers or ex-police officers themselves. For example, in Mexico City, much of the current crime is actually committed by these law enforcement agents.
What caused these ex-police officers to change? What they worked to prevent before, now they are guilty of. We need to identify these type of "marginalized" groups of individuals, including the unemployed and school dropouts, as potential offenders. While increased unemployment and dropout rates do not cause the increase in crime rate, it is one of the factors that may contribute to this outcome.
In Venezuela, democratization is linked to change because there was essentially a shutting down of the police within poorer areas. After dark, not a single police officer is seen in the smaller, poorer towns. Venezuela seems to identify two types of citizens: "first and second class" citizens, and only the first-class benefit from the country's Stategic Plan.
In Buenos Aires, where guns are cheaply bought on the black market, 62 percent of citizens have been or have family members who have been robbed within the past six months.
Some people believe the rising crime rates in Latin America will peak and then naturally fall again. However, recent crime activity points more towards an epidemic, causing permanent change. Globally, we are seeing permanent social change and a product of this change is the rising crime rates. We must analyze the reasons such social change is occurring: Is it due to poverty and economic factors? Is it due to cultural and family factors, such as destructive video games watched by children? All of these theories should be included when analyzing the issue of increased crime.
1 Summary by Corinne Davis.
2 Summary by Peter Ward. Although this is included in the main body of the Workshop Memoria, the organizers wish to expres their gratitude to Dr. Briceño-León for agreeing at short notice to offer his paper as the plenary address on Thursday 6th May.
3 This paper arrived while the Memoria was in production - hence it is included unpaginated. It was edited for length by Peter Ward. Please contact Dr Dellasoppa for a copy of the full paper and for further details of the research project
4 Data from the Ministry of Health.
5 International survey by the United Nations on the use of firearms. Also, Folha de São Paulo, May 6th, 1997, p. 3-7. Considering the country at a global level, these same indicators in other countries remain at similar levels. For example, in the United States the homicide rate for every 100 thousand inhabitants evolved from 9.9 in 1979 to 9.4 in 1994.
6 Arriaga, E. 1984. Measuring and Explaining the Change of Life Expectancies. Demography, 21(1), 83-96.
7 It should be pointed out that there is probably a certain under par registering of AIDS cases, because the doctor filling out the death certificate doesn't always put the syndrome as being the cause of death, but rather it's opportunist manifestations or infections.
8 This paper was presented a few weeks ago in Buenos Aires for a public which is not acquainted with the criminology literature. Therefore, the people in this workshop may find some observations very obvious and simplistic. My goal at this roundtable was to call for better data and some reflections on crime theory, which in Argentina remains very outdated.