Neighbourhood characteristics and the distribution of crime in Halifax

View the most recent version.

Findings >

Halifax in context
Distribution of crime in the City of Halifax in 2001
Neighbourhood characteristics and crime
Descriptive results: a comparison of high- and lower-crime neighbourhoods
Results of multivariate analysis
Summary of findings – Halifax

Halifax in context

In 2001, the Halifax census metropolitan area (CMA), with a population of 359,183, ranked thirteenth in terms of size among the 27 CMAs in Canada.¹ The 2001 population was up 4.7% from 1996 (342,851), while the population for Nova Scotia remained virtually unchanged over the same period. In 2001, the Halifax CMA represented approximately 40% of the population of the province of Nova Scotia. By the 2006 Census, the population of the Halifax CMA had increased 3.8%, reaching 372,858 residents.

For many years, the area now known as the Halifax Regional Municipality existed as four separate municipalities. They were Halifax, on the south-west side of the harbour, Dartmouth on the north-east side of the harbour, Bedford at the top of Bedford Basin, and Halifax County, which makes up the remainder. These four areas amalgamated in 1996, to be governed by a single city council.

In 2001, the Halifax Regional Municipality was serviced by two police detachments. The Halifax Regional Police was responsible for policing the urban core of the municipality, which was divided into three divisional areas: Peninsular Halifax (Central Division), Bedford to Sambro Loop (West Division) and Dartmouth (East Division). These three areas accounted for 56% of the municipality. The remaining 44% was serviced by the Halifax County Rural RCMP detachment and was made up of the following areas: Sheet Harbour, Musquodoboit Harbour, Cole Harbour, Tantallon and Lower Sackville. This study focuses on the portion of the Halifax Regional Municipality that was policed by the Halifax Regional Police,² which covers approximately 160 square kilometres divided into 51 census tracts (CTs),³ and had a population of 191,514 in 2001 (Map 2.1).⁴

Chart 2.1 compares Halifax's crime rate to other major CMAs as well as the overall rate in Canada from 1991 to 2006. The Halifax CMA follows the general trend of declining crime rates in Canada although the CMA's rate has consistently hovered above the national average throughout the 1990s and early 2000s. In fact, victimization data from the 2004 General Social Survey indicates that the total violent victimization rate reported by residents of the Halifax CMA (229 per 1,000 inhabitants aged 15 and over) was more than double the Canadian average (106 per 1,000 inhabitants aged 15 and over). However, no statistical difference was found between the rate of household crime in Halifax and the national average, 293 incidents of household crime per 1,000 households and 243 incidents of household crime per 1,000 household, respectively (Gannon and Mihorean 2005).

Offence categories included in this study are violent, property, drug, prostitution, offensive weapons, and gaming and betting offences. In 2001, the Halifax Regional Municipal Police reported nearly 21,000 of these selected offences, the vast majority of which were property crimes (77%), followed by violent offences (20%) and other offences (3%) including prostitution, drug-related offences, offensive weapons-related crimes and gaming and betting offences. This distribution of offences is relatively similar to the crime composition for Canada overall (79%, 17% and 4%, respectively).

Distribution of crime in the City of Halifax in 2001

Whereas incidents were reported in 50 of the 51 CTs that make up the area policed by the Halifax Regional Police Service (Table 2.1 and Table 2.2), a closer look at the distribution of incidents reveals that police-reported crime is not evenly distributed across the city but rather clustered in certain areas. Map 2.2 and Map 2.3 show violent and property crime hot spots in the city of Halifax area in 2001. Clusters of violent crime (Map 2.2) appear in the downtown area located near the harbour, as well as across the harbour to the east, in the area previously called the city of Dartmouth. Warm spots of violent crime also appear scattered throughout the city in various locations. The distribution of property crime incidents in Halifax (Map 2.3) is quite similar to the distribution of violent crime. One exception is two additional hot spots east of Halifax Harbour, located near two busy shopping centres.

The spatial distribution of crime in 2003 was approximately the same as in 2001. The increase in property crime was distributed among the same hot spots, and the concentrations of violent crime were almost identical. The data correlation coefficients at the CT level of 2001 and those of 2003 show the similarities; the correlation coefficient is 96% (p<0.001) for violent crime and 95% (p<0.001) for property crime. The crime data geocoded in 2001 seem to represent a definite general tendency of crime distribution in the city of Halifax area (Map 2.17 and Map 2.18 in "Appendix 2: Neighbourhood characteristics and the distribution of crime in Halifax").

It is important, however, to consider not just the relative distribution of crime across a city, but also to take the city's population into consideration. Map 2.4 and Map 2.5 illustrate crime distribution taking into account population at risk.⁵ When Halifax's population at risk is included, many of the hot spots disappear and violent crime hot spots remain on either side of Halifax harbour, with similar results for property crime. Property crime warm spots continue to appear east of Halifax Harbour in the areas of the large shopping centres, as well.

Neighbourhood characteristics and crime

This section explores the relationship between demographic, socio-economic, and land-use characteristics and 2001 rates of violent crime and property crime by population at risk in Halifax neighbourhoods.⁶ The analysis makes use of total violent and property crime rates rather than looking at individual rates of specific offences in order to maximize the number of incidents being considered. It should not be concluded from the results of this study that some neighbourhood characteristics are the cause of crime; rather the results show that these factors are associated or co-occur with higher crime rates in neighbourhoods.

Descriptive results: a comparison of high- and lower crime neighbourhoods

To examine the relationship between violent and property crime rates and selected neighbourhood characteristics, the 51 CTs have been divided into two groups for each crime type. The first group contains CTs falling into the highest 25% of property and violent crime rates, and the second group contains the remaining 75% of CTs. The differences are significant at p<0.001 unless otherwise indicated, based on an independent samples T-test.

Before controlling for other factors being studied, significant differences are noted in selected characteristics when comparing neighbourhoods with high crime rates to those with lower crime rates. For instance, when examining population characteristics, it can be seen that CTs with the highest rates of violent crime have significantly greater residential mobility, as indicated by the proportion of people who moved in the year preceding the census, compared to lower violent crime CTs (26% and 20% respectively) (Chart 2.2). The highest violent crime rate CTs also have significantly greater proportions of lone-mother families (23% and 15%, respectively), people living alone (22% and 15%, respectively), and people living in common-law situations (11% and 8%, respectively) than do their lower violent crime rate counterparts. There was no significant difference between high and lower crime neighbourhoods in terms of the proportion of young males living there.

Neighbourhoods with the highest property crime rates also display significant differences when compared to their lower property crime rate counterparts in terms of the proportion of people living alone or living in common-law situations (Chart 2.3).

An examination of socio-economic characteristics also highlights a number of significant differences between the highest crime rate and lower crime rate CTs. The most significant differences are seen between the highest violent crime rate CTs and their lower violent crime rate counterpart (Chart 2.4). The highest violent crime rate CTs have a higher unemployment rate (9% and 7%, respectively), a higher proportion of people below the low income cut-off (30% and 18%, respectively) and a higher proportion of total income from government transfers (17% and 12%, respectively). They also have significantly lower median household incomes ($33,000 and $47,000, respectively) than their lower violent crime rate counterparts. Fewer significant differences in socio-economic characteristics were seen between the highest property crime rate CTs and lower property crime rate CTs, as seen in Chart 2.5.

Finally, Chart 2.6 and Chart 2.7 demonstrate differences between high-crime and lower-crime neighbourhoods in terms of land use and housing characteristics. In high violent crime areas, a slightly greater proportion of housing was in need of major repair (10%, and 7% in low violent crime areas, respectively), and a lower proportion was owner-occupied (15% and 22%, respectively) than in the lower violent crime rate CTs. Also, in high violent crime areas, a higher percentage of households spent more than 30% of their income on shelter (19% and 13%, respectively). The highest violent crime rate neighbourhoods had clearly greater proportions of commercial zoning.

There were fewer statistically significant differences in land-use and housing characteristics between the high and lower property crime areas. The highest property crime CTs had statistically significant greater proportions of households spending more than 30% of their revenue on housing (18% and 13%) and a greater amount of commercial zoning (20% and 7%) than the lower property crime CTs.

Results of multivariate analysis

When considered individually, the neighbourhood characteristics discussed above are associated with violent and property crime, with some characteristics occurring in neighbourhoods with higher crime rates, and others occurring in neighbourhoods with lower crime rates. The strength of these associations can vary when multiple characteristics are considered together.

The presence of Halifax Harbour separates the city into two unique areas. To determine whether or not this would have an effect on the models, a location variable was introduced to indicate whether each CT fell into the area north-east of the harbour (formerly known as Dartmouth) or the area south-west of the harbour.⁷ In both the property crime and violent crime models, this location variable was significant, indicating that a neighbourhood's location, either in the north-east or south-west area of the city, makes a difference to its crime rate. To determine if different factors are associated with crime rates in these two areas of the city, separate regression models were run for each area of the city.

Results of the regression models are shown in Table 2.3 and Table 2.4. The violent crime model for the north-east area of the city includes 3 explanatory variables, and results in an Adjusted R-Square value of 0.80, indicating that the model explains approximately 80% of the variation in violent crime rates in those neighbourhoods. The estimated regression coefficients (b) provide an indication of the relative contribution of each variable after controlling for the other variables in the model.

The crime rate regression model shows that the proportion of the population aged 20 and over with a bachelor's degree makes the largest relative contribution to the explanation of violent crime in the north-east area (b=-0.48, p<0.01). Therefore, as the proportion of the population holding a bachelor's degree increases in a neighbourhood, the violent crime rate decreases. This characteristic seems to offer protection against crime at the neighbourhood level. In contrast, the rate of violent crime is higher in areas with higher proportions of lone-parent mother families (b=0.41, p<0.05). The proportion of commercial zoning (b=0.32, p<0.001 is another factor that helps explain the variation in violent crime rates in the north-east neighbourhoods of Halifax. The greater the proportion of commercial zoning, the higher the violent crime rate in that neighbourhood.

The variables affecting violent crime rates in the south-west area of the city differ from those in the north-east. The violent crime model for the south-west includes only 3 explanatory variables, and results in an Adjusted R-Square value of 0.60, indicating that the model explains approximately 60% of the variation in violent crime rates in those neighbourhoods. The proportion of the neighbourhood population living alone makes the largest relative contribution to the explanation of violent crime rates in the south-west of the city (b=0.91, p<0.001). The next largest contribution is the proportion of lone-parent mother families (b=0.33, p<0.01), followed by the proportion of dwellings in need of major repair (b=0.30, p<0.05). As each of these factors increase in Halifax's south-west neighbourhoods, so do their violent crime rates.

The results of the property crime regression models also indicate a difference in which factors contribute to higher property crime rates on either side of the harbour. The property crime model for the north-east area includes two explanatory variables, and results in an Adjusted R-square value of 0.42. The largest relative contribution to the explanation of property crime rates in the north-east area of the city is made by the percentage of commercial zoning in a neighbourhood (b=0.33, p<0.05). The more commercial zoning a neighbourhood has, the higher the property crime rates. Neighbourhoods with higher proportions of people that are unemployed also have higher property crime rates (b=0.30, p<0.05).

The property crime regression model for the city's south-west area explains slightly more of the variation in property crime rates (47%) than the model for the north-east neighbourhoods, as indicated by an Adjusted R-square value of 0.47. The proportion of unaffordable shelter, as indicated by the proportion of the population spending more than 30% of their income on shelter, makes a significant contribution to the explanation of property crime rates in the south-west area of the city (b=1.36, p<0.001). However, the other variable making a significant contribution to the explanation of the variation in crime rates in this part of the city is median household income (b=0.60, p<0.05). As the median household income in a neighbourhood increases, so do property crime rates. While having these two income-related measures being significant in the same model appears somewhat contradictory, that isn't necessarily the case. It may be that some households in Halifax with relatively high incomes spend more than 30% of that income on shelter.

Summary of findings – Halifax

In Halifax, property and violent crime hot spots were located largely in the city's downtown area and east of Halifax Harbour. When all other neighbourhood characteristics are taken into account, results indicate that the factors linked to the variation in neighbourhood crime rates based on the population at risk are different in areas north-east of Halifax Harbour than they are in the areas south-west of the harbour.

In fact, violent crime rates on either side of the harbour are higher in neighbourhoods with more single-mother families. These families tend to be living in low-income situations. Violent crime rates north-east of the harbour also occur in neighbourhoods with larger proportions of commercial zoning and populations with lower levels of education. In the area south-west of the harbour, violent crime rates occur in neighbourhoods where more people live alone, and the housing situation is poor, as indicated by the proportion of houses in need of major repairs.

Property crime rates in the north-east area are higher in neighbourhoods with more commercial zoning and higher rates of unemployment. On the south-west side of the harbour, a neighbourhood's property crime rate increases with higher proportions of households spending more than 30% of their income on housing, as well as higher median household incomes.

Footnotes

1. These population figures are based on the Statistics Canada census metropolitan area (CMA) population for Halifax, and are used for national comparative purposes. The CMA boundary includes adjacent municipalities situated around the urban core and is consequently larger than the study area.
2. A small portion of the area services by the Halifax Regional Police is excluded from this study because of a discrepancy between police service and Census Tract boundaries. The residential population of this area is 6,380 people, which is less than 3% of the population policed by the Halifax Regional Police Service.
3. The neighbourhoods used in this analysis correspond to Census Tracts (CTs), which are small, relatively stable geographic units that usually have residential populations between 2,500 and 8,000 people, and are located within census metropolitan areas.
4. Halifax Regional Municipality. 2006. "Halifax Regional Police" www.halifax.ca/Police/PatrolAreas/index.html
5. To obtain more detailed information on the use of the population at risk in the spatial crime data analysis, see the Methodology section.
6. The definition of the term 'neighbourhood' used in this study reflects the census tracts (CTs). For additional information, see the Methodology section.
7. The city was divided into a North-East section made up of 19 CTs, and a South-West section made up of the remaining 32 CTs.