DEBATE DEBATE
Debate on the paper by Roberto Briceño-León
Debate sobre el artículo de Roberto Briceño-León
Juan Mario Fandino-Marino
Instituto de Filosofia e Ciências Humanas, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brasil. fandinom@uol.com.br
The article provides a wealth of statistical information on homicides in Latin America and on some of the correlates traditionally discussed in the criminological literature. This information originally appears rather chaotically in different sources, and the author presents them here in an organized way, which has great merit for the debate on the issue and represents a huge research effort. Among the correlates are poverty, level of schooling, employment, family structure, and urbanization. Other traditional correlates, especially youth aspirations and religion, are treated at a more conceptual level. Even so, their treatment systematizes ideas which are not necessarily original, but which form a valuable theoretical framework.
Another highly traditional quality, yet still noteworthy in the overall context of Latin American sociology, is the fact that the article is based on a dialogue between theory and data, which naturally launches a disciplined debate. It is thus a source of satisfaction to encounter such a paper for these reasons. Following these remarks, we now approach our critical sociological position towards the study.
As for the simple information on homicides, a very important recent trend, already recognized in the literature, escapes the battery of data provided by the author, namely a highly significant drop in homicides in some key cities, like São Paulo (37.0% in five years), Bogotá (from 80 to 23 per 100 thousand in seven years), and Cali (25.0% in nine years) (Kahn T, Zanetic A. Personal communication; 2005). These decreases are relevant for the issue, since they may result both from successful public policies (Kahn T, Zanetic A. Personal communication; 2005), as well as cyclical historical trends 1,2, or both, and they find no echo in the argument presented by Briceño-León.
The discussion on cities and urbanization provides theoretical elements that are not mistaken, but which are far from constituting a basis for what the author calls a "sociological framework for the explanation of violence". The discussion at stake is limited to the assertion (as evident as it is un-analytical) that the cities are not what we wish they were, and are what we wish they were not. The affirmation that "Latin America cites were a place of hope for security and law, hence the great rural-to-urban exodus in the 1940s and 50s" simply fails to agree with the data 3. In the latter study we find that in Colombia the attraction exerted by cities is equivalent to only 10.0% of the explanatory power of expulsion by demographic pressure and agricultural technology, the most important factors at the time. The etiological literature on migration generally did in fact contemplate the security issue, but in a localized way and in specific cases, and it has generally never been considered a determinant factor in Latin America.
The problem with the author's structuring of a sociological framework in three causally different dimensions (macro-factors that "originate", meso-factors that "foment", and micro-factors that "facilitate" violence), begins when he abandons the causal epistemological status proper to each of the three traditional levels of analysis for crime: micro, meso, and macro. The traditional effort in the literature working with causal links between these levels is very clear, beginning with Sutherland 4 in 1924, moving on to Cloward & Olhin 5 in 1960, and reaching the "integrated models" in recent years (where the explicit objective of analysis is to work with this issue). This is what is at stake in terms of an epistemological framework, and not the three ad hoc dimensions of Bricenõ-León. It is true that the "violentological" and criminological specificity of Latin America should be understood at the macro level, which would justify placing this specificity as having originated at this level of aggregation. But contemporary urban violence in Latin America is not such a unique legacy for our continent, as shown by Gómez-Buendia 6. The links are much more complex.
As for the author's empirical analysis, it is generally limited to presenting data on each aspect or variable separately, after which, based only on their approximate contemporaneity, (!?), he derives conclusions on their causal nexus with homicides. In other words, the author's empirical analysis contains practically no relational factual evidence where we might observe some types of co-variations or associations. The case in which the author approaches a relational empirical methodology is his analysis on the relationship between poverty, urbanization, and homicide rates. We do not refute that in a broad and generic sense, these phenomena and their interactivity (correctly emphasized by Briceño-León) may be associated in some way. However, at least based on the Brazilian studies on the subject, we know that the relationship between poverty and homicides is not linear, and that extreme poverty levels, including those in urban areas, are not the ones that stand out as factors for homicides. Following the level of spatial-temporal aggregation that Briceño-León intends to adopt, rightfully and pertinently, the treatment of the hypotheses raised is virtually worthless, methodologically speaking. In fact, based on a simple visual inspection of Table 6, invoked by the author, one cannot conclude in favor of the rigor of his hypothesis. Based on a superficial test using the author's own Table 6, these relationships, as I will illustrate next, prove to have very little explanatory power, even though they may be interesting. Based on the data from Table 6, we calculated the following multiple regression equation:
H = a+b1X1+b2X2+b3X 1X2+E,
were H represent the homicide rates and X1 and X2 represent poverty and urbanization, respectively. Despite the calculating problems based on n = 16 (very small), the results are: a precarious adjusted R2 of 0.074; the betas, even with the model's precarious overall adjustment, are interesting and lend some credit to Briceño-León's theory: -1.67 for poverty, -0.76 for urbanization (both negative!), and finally a positive beta of +1.59 for the multiplicative interactivity term. None of the coefficients is significant at 0.05.
1. Fandino-Marino JM. A violência na América Latina e seus ciclos altruísta e egoísta/anômico. Revista do Direito 2000; (14):1-22.
2. Fandino-Marino JM. The moral cycle of egoistic and altruistic violence: a century of bloodshed in Colombia. In: Anderson M, editor. Cultural shaping of violence. West Lafayette: Purdue University Press; 2004. p. 268-83.
3. Fandino-Marino JM. Determinantes econômicos e sociológicos da migração rural-urbana. Revista de Economia Rural 1979; 6:124-42.
4. Sutherland E, Cressey D. Principles of criminology. Chicago: J.B. Lippincott; 1955.
5. Cloward R, Ohlin LE. Delinquency and opportunity. New York: Free Press; 1960.
6. Gómez-Buendia H. Urban crime: global trends and policies. Tokyo: United Nations University; 1989.