Crime is a complex phenomenon that occurs when an offender, a victim and a law intersect in time and space. Crimes, whether they be property crime, violent crime, white collar crime or nuisance crime, are prevalent in all societies with many costs that are both monetary (property loss, insurance, law enforcement, the judiciary and corrections) and psychological (victimization and safety) (Sharpe 2000). Though different crimes have differential effects on the monetary and psychological costs to society, the reduction of crime, in general, has a positive effect on society.
Within the literature, social disorganization theory and routine activity theory have frequently been employed to analyze crime geographies. These theories respectively focus on issues of social deprivation and routine activities as predictors of the spatial pattern of crimes rates. This article synthesizes these theories as the basis for a spatial autocorrelation analysis of crime rates in the Vancouver census metropolitan area (CMA). The Vancouver CMA is the third largest metropolitan area in Canada (2 million population) and the largest metropolitan area in western Canada; the City of Vancouver, the geographic area under study, had a population of 514,000 in 1996. Vancouver has also experienced recent substantial growth in its population, by 8.5 percent from 1996 to 2001. Information is drawn from the calls for service (CFS) made to the Vancouver Police Department (VPD) along with socio-demographic and socio-economic data from Statistics Canada's Census data.
Within the Canadian context, crime rates in Vancouver are higher than expected on the basis of population size. Thus, with respect to the rate for all criminal code infractions, not considering traffic offences, the Vancouver CMA (1996) is first in Canada (14,868 per 100,000 persons), more than doubling the rate found in Toronto (6,920 per 100,000 persons) and almost doubling that in Montreal (7,979 per 100,000 persons). The same relative standing holds for the property crime rate in Vancouver (10,494 per 100,000 persons) in comparison to Toronto (4,277 per 100,000 persons), and Montreal (5,351 per 100,000 persons), and to a lesser extent for the violent crime rate in Vancouver (1,325 per 100,000 persons) in comparison to Toronto (824 per 100,000 persons), and Montreal (839 per 100,000). These differences in crime rates, however, have been decreasing in recent years (Kong 1997; Wallace 2003).
This article differentiates itself from the previous literature on the geography of crime in four ways. First, Vancouver crime data have not been analyzed before at such a disaggregated spatial scale. Second, the crime data and the census data correspond to the same year. This correlation is particularly lacking, due to data availability, in the United States where the census is only performed every 10 years. The lack of correspondence between crime data and census data is a difficulty when attempting to assess causal relationships, unless one can assume neighbourhood stability in all characteristics used in the analysis. Third, an explicitly spatial analytical technique, spatial autoregression, is employed in the analysis, in contrast to much inferential work on the geography of crime that favours classical statistical analysis. Fourth, three types of crime (violent crime, break and enter, and automotive theft) are investigated separately to assess the different demographic, economic, and social attributes of census boundary units within Vancouver. Because the relationships between the set of independent variables and the different crime rates are allowed to vary, the analysis is provided with greater flexibility in assessing the predictive power of social disorganization theory and routine activity theory for different crimes in the Vancouver CMA.
The following section reviews the literature on social disorganization theory and routine activity theory in the geography of crime. …