Sampson concludes that “the influence of SES on police contacts is contextual in nature, and stems from an ecological bias with regard to police control” . Whether economic deprivation has a direct relationship with crime through individuals’ behavioral mechanisms , or an indirect effect on crime rates through the operation of police bias, there is a strong theoretical basis for the finding that poor neighborhoods report higher crime rates than rich ones. The measures of economic deprivation that are controlled for in this study are poverty and unemployment. Mollie Orshansky developed the original poverty thresholds in 1963-64 when she was an economist working for the Social Security Administration, shortly before the declaration of a “War On Poverty” by President Johnson . An individual or household is in poverty when its total cash income falls below the applicable threshold, determined by family size and composition . The 2010 poverty thresholds range from $11,139 for a single individual living alone to $42,156 for a family of eight or more people living in the same household. Indicators of socioeconomic disadvantage, including poverty and unemployment, have been associated with higher crime rates in Miami, Florida and Columbus, Ohio . Other studies show that rates of crime and violence are extremely high in neighborhoods containing public housing developments, which are areas of extremely concentrated socioeconomic disadvantage . In their study of medical marijuana dispensaries in Sacramento, Williams and colleagues found that violent crimes were significantly associated with “concentrated disadvantage”, a variable constructed from 2008 poverty guidelines. Property and violent crimes were not associated with density of marijuana dispensaries, indoor weed growing accessories but both categories were significantly related to unemployment rate . Further evidence of the link between crime and unemployment is found in crime data from across the United States in the 1990’s, a decade of incredible crime reduction.
Raphael and Winter-Ebmer found that a substantial amount of the reduction in property crimes could be explained by the corresponding decline in unemployment rate. A weaker relationship existed between unemployment and violent crime, according to the researchers. Freeman presents similar findings and concludes that as much as one-third of the drop in crime in the 1990’s can be explained by the expanding job market. Theorizing from earlier empirical work by Sampson , which found that macrolevel indicators of family disruption were related to rates of juvenile crime, Sampson and Groves include family disruption among their “exogenous sources of social disorganization” . The theoretical basis for this lies in the notion that “traditional” families provide their communities with greater parental supervisory resources, compared to single-parent families. Additional supervision results in greater social control and more effective prevention against crime . This theoretical link is supported by empirical evidence. In their reassessment of Sampson and Groves’s original analysis, Veysey and Messner report that family disruption has significant relationship with crime even independently of other social disorganization processes. Indeed, percent of single person households was significantly related with crime in the Sacramento dispensary studies conducted by Williams and colleagues .According to social disorganization theory, communities with higher levels of residential turnover suffer from correspondingly lower levels of social control and are therefore likely to report higher rates of crime . The present study conceptualizes “residential instability” as an index of the percent of housing units in a given census tract that are vacant and the percent of individuals living within the tract who are between the ages of 18 and 29. These variables are also used by Martínez and colleagues in their study of crime and drug use in Miami neighborhoods. Williams and colleagues found that property crimes—but not violent crimes— were significantly associated with percent of owner-occupied households, which is a measure of residential stability. Hipp and colleagues examined residential turnover in an ethnic context and found it to be significantly associated with crime. They conceptualized “residential stability” as the average length of residence of households in the relevant census tract .
Although it is not included in the forthcoming analysis, there is one final measure of social disorganization that appears in the theoretical literature which is relevant to the present discussion of crime in city neighborhoods: population heterogeneity . The argument here is that segregated communities suffer from lower rates of communication and interaction, which prevent them from organizing collectively to reduce crime and delinquency—even when the different population groups have a shared interest in law and order. A number of studies have found that population heterogeneity is associated with higher crime rates . Conclusion In this chapter I have reviewed the literature on cannabis, MCDs, and crime that is relevant to the present study. Particular attention has been given to routine activities theory as well as social disorganization theory . In the next chapter I extrapolate from these theories in developing a conceptual model of crime to test for the criminogenic effect of MCDs.Empirical testing is needed in order to assess the strength of these claims. I develop the model that will be used to test these claims in the first section of this chapter. In the latter sections I describe the various measures that comprise this model, which are tested in the following chapter. In comparing crime rates across city neighborhoods, it is important to control for neighborhood characteristics that are related to crime independently of MCDs. In the present study I consider three vectors of social and demographic variables, drawn from social disorganization theory, that will be used as control variables in my model: economic deprivation, family disruption, and residential instability . “Social disorganization” is a concept used to describe communities in which there is a lack of sufficient social cohesion among members—characterized by low socioeconomic status, fewer stable families, and high residential turnover—that would otherwise serve as a deterrent against crime. Higher levels of these social disorganization indicators are associated with higher crime rates at the citywide and neighborhood level . Controlling for these variables across city neighborhoods will help determine whether, and to what extent, MCDs have an independent effect on crime. Current debates about the relationship between MCDs and crime tend to revolve around politicized rhetoric and anecdotal claims .
The only scholarly discussion of this question known by this author at the time of this writing consists of a pair of studies out of UCLA that examined crime data from 95 census tracts in Sacramento for the year 2009 . Nancy J. Williams and colleagues found that crime rates were not significantly associated with neighborhood density of medical marijuana dispensaries. The researchers identified several other variables with higher explanatory power with respect to crime, including unemployment, percent of the population that is young, and percent of one-person households. These findings correspond with the assertion made by social disorganization researchers that low levels of SES, high residential instability, and high rates of family disruption correspond with relatively higher rates of crime. In the following sections of this chapter, I will discuss these and other measures that comprise the empirical model developed by this study. First, I will briefly review the two criminological theories that I draw upon: routine activities theory and social disorganization theory.Stemming from human ecology, routine activities theory attempts to explain crime on the basis of three conjoining factors: suitable targets, likely offenders, and absence of capable guardians . This approach suggests that MCDs could potentially lead to crime by increasing the concentration of likely offenders in the local neighborhood, or, more directly, by themselves presenting suitable targets for crime. These are precisely the sorts of claims that are made by critics of MCDs . Conversely, it is conceivable that MCDs might decrease crime in a neighborhood by introducing increased guardianship, rolling benches which could serve as a deterrent against crime. Proponents of MCDs make arguments of this type when asserting that dispensaries—especially those in compliance with local regulatory schemes requiring stringent security protocols—actually reduce crime in their neighborhoods . Routine activities theory is an appropriate and compelling lens through which to analyze these competing claims. It is appropriate because it presents direct causal mechanisms through which MCDs might affect crime, and it is compelling because those mechanisms plausibly run in either direction . Another useful approach found in the criminological literature is social disorganization theory, first developed by Shaw and McKay, which links urban crime to the concept of “social disorganization” . Studies have identified several measures of social disorganization that are positively associated with crime . In the present study I examine three “exogenous sources of social disorganization”: socioeconomic deprivation, residential instability, and family disruption. If these neighborhood characteristics predict crime, as social disorganization theory suggests, then they ought to be incorporated into more specific models of urban crime—including the present attempt to explain the relationship between MCDs and crime. I do not argue that there is a direct relationship between MCDs and social disorganization. Rather, I look to social disorganization theory to identify neighborhood characteristics, independent of cannabis, which might confound the real relationship between MCDs and crime. Variables from each of the three categories of social disorganization variables used in the present model are summarized in Table 3.1 below. For a more thorough review of the underlying theories and concepts, see the previous chapter.This study examines the spatial relationship between medical cannabis dispensaries and crime across 189 census tracts in San Francisco in the year 2010. I test two hypotheses drawn from the theoretical literature on routine activities theory , controlling for neighborhood characteristics drawn from social disorganization theory . The first is that MCDs increase crime by attracting likely offenders and presenting them with suitable targets; the second is that MCDs actually decrease crime by protecting their surrounding community with adequate security measures and thereby providing capable guardianship.
The theoretical bases for these claims are discussed in greater length in the previous two chapters. In this chapter I present the research methodology employed by this study and discuss their implications. Data are collected from the San Francisco Police Department, Planning Department, and Department of Public health; the California Department of Finance; the American Community Survey ; and the United States Census Bureau. Linear regression models are tested using four dependent variables at the census tract level: total property crimes, property crimes per 1,000 residents, total violent crimes, and violent crimes per 1,000 residents. Findings are largely but not perfectly consistent across these different models with respect to the spatial relationships between crime, MCD density, and the seven other neighborhood characteristics analyzed: poverty, unemployment, family stability, vacancy rate, percent of the population ages 18-29, percent of the population that is male, total population size, and percent of land commercially zoned. Findings indicate a consistently weak but statistically significant relationship between MCDs and crime. MCD-containing tracts have slightly higher rates of both property crime and violent crime than tracts that do not contain MCDs. The link is stronger for violent crime than property crime . I am careful not to infer too much from this finding, due to the limited number of cases under review—26 dispensaries across 16 census tracts, compared to 173 non-MCD-containing tracts. Crime is more strongly predicted by the social disorganization variables examined by this study: socioeconomic disadvantage, family disruption, and residential instability. Consistent with social disorganization theory, socioeconomic disadvantage and family disruption are found to be strong predictors of high crime rates. By “stronger” I mean that it has a larger correlation coefficient and a higher degree of statistical significance . “Family stability” is negatively associated with crime across all four models. Again, this link is stronger than the link between MCD density and crime. The third category of variables drawn from Sampson and Groves’ “exogenous sources of social disorganization” , residential instability, was not as strongly predictive of crime as socioeconomic disadvantage or family disruption. In the following sections I explore these findings more deeply and discuss their implications for research as well as policy making. In a recent working paper for the California Center for Population Research at UCLA, Nancy Williams and colleagues present a routine activities approach for examining the link between MCDs and crime. They examine 95 census tracts in Sacramento using data for the year 2009. Their findings indicate that tracts containing dispensaries are not significantly associated with higher rates of crime when controlling for neighborhood characteristics associated with crime . This study borrows from their work in conducting a cross-sectional analysis of the spatial relationship between MCDs and crime in 189 San Francisco census tracts for the year 2010. All measures are aggregated to the level of census tracts.