Abstract: Canadian homicide rates have declined steadily since the mid-1970s, though this overall trend has been punctuated by temporal and regional fluctuations. It is almost consistently noted that young males are overrepresented in the interpersonal violence equation, and changes in this demographic can greatly affect homicide rates. Yet the ubiquity of the positive effect of age-composition has been questioned. Using fixed-effects analysis, this paper examines the relationship between young males and homicide rate changes over a thirty-year period. Results indicate that homicide rate changes in Canada are indeed a function of changing demographics; however, the relationship is complex, and socio-economic factors both mitigate and exacerbate this relationship.
Keywords: demographics, Canada, fixed-effects, homicide
Over the long term, western societies have experienced a decrease in levels of interpersonal violence (Elias  1978; Gurr 1981),1 but more recent homicide patterns demonstrate geographical and temporal fluctuations (Archer and Gartner 1984; Blumstein and Rosenfeld 1998). The causes and correlates of homicide are complex and remain an important focus of debate for criminologists: some argue that sociological, economic, and cultural factors all affect homicide patterns (Phillips 2006). One of the most consistent explanations for the uneven distributions of violence over time and place is demographic variation (Andresen et al. 2003; Blumstein 2006; Cohen and Land 1987; Fox 2006; Fox and Piquero 2003; Gottfredson and Hirschi 1990). Slight changes in the population age structure over time are evident, and it has been well documented that young males are disproportionately responsible for criminal activity (Gottfredson and Hirschi 1990).
Gottfredson and Hirschi (1990) maintain that the age-crime relationship holds across time and location, and research at the individual level demonstrates a strong association between age and crime (Blumstein et al. 1986; Cohen and Land 1987; Fox and Piquero 2003; Hirschi and Gottfredson 1983). There is also support for a causal relationship between age composition and rates of interpersonal violence at the aggregate level (Fox 2006; Fox and Piquero 2003). However, there is significant disagreement regarding whether demographics actually dictate the homicide rate, or if other factors of equal, or perhaps more, weight might affect change in rates of violence (Blumstein 2006; Fox 2006; Gartner 1990; Gartner and Parker 1990). For example, research findings related to age composition overall are inconsistent, in that they sometimes indicate an interaction between demographics and changing socio-economic factors (Blumstein 2006; Pampel and Gartner 1995; Phillips 2006).
The relationship between age composition and violence is largely supported by national studies, but it has less support cross-nationally (Gartner 1990). In addition, although demographics are widely accepted as a strong predictor of interpersonal violence rates in the United States, it is also recognized that they influence Canadian rates even more (Andresen et al. 2003). Canada is a particularly fertile field within which to examine changing rates of homicide given the regional variation in demographics, population density, heterogeneity, and economic factors. National examinations of homicide in Canada have not thus far studied the interaction between demographics and socio-economic factors. Examining how the correlates and causes of homicide function can illuminate certain aspects of interpersonal violence and social structure in Canada. In order to examine Canadian homicide rate fluctuations and the factors which influence temporal/geographical trends of violence, this analysis utilizes the varying homicide rates in seven regions from 1976 to 2005. A time series cross sectional (TSCS) dataset was constructed using information from both the Canadian Homicide Survey and Statistics Canada resources, and a fixed-effects analysis was utilized to examine the effects of structural factors on homicide rates. …