Table 3 compares the profile of the three classes of tracts. The first column describes the 1,933 other tracts, the next column the 155, potentially gentrifying, and the third the 94 gentrifying tracts. The comparison is examines the following factors:
Rent and Income Levels: Increases in average incomes and rents between 1981 and 2001 provide an indication the tract may be gentrifying.
Distance from the CBD: The straight-line distance from the centre of the census tract to downtown measures the centrality of the census tract to amenities that are commonly sought after by gentrifying households.
Population and Dwelling Density: While neighbourhood density changes were not commonly reported in previous studies of gentrification (Beauregard 1990, Helms 2003, Ley 1988, 1993), density changes might indicate the presence of conversions, redevelopment and a shift to smaller households.
Dwelling Type, Tenure and Age: Changes in the proportion of single-detached, high-rise apartments and other dwellings as well as the proportion of owned and rented dwellings per census tract describe some of the likely attributes of gentrification.
Demographic Characteristics: Typically gentrifying households are young, well-educated, highly mobile and single-person households. Thus, the census variables that measure the proportion of single-person households, persons per households, movers (i.e., persons who changed their residential address since the previous census), and proportion of persons aged 25 to 39, and percentage of persons with university degrees, helps to develop a profile of the census tracts that might be undergoing gentrification (Ley 1988, Wyly and Hammel 1999).
Rent and Income Differences
All three groups of tracts increased in average rents between 1981 and 2001; however, the gentrifying tracts had the largest absolute increase in average rent.The average household income was increasing much faster in gentrifying tracts than in the other two categories. For example, average personal income in 1981 was just under $20,000 for the gentrifying tracts; this is slightly lower than the reported average income for remaining potentially gentrifying tracts. In 2001, however, average incomes increased to almost $26,000 in the gentrifying tracts, much higher than the $23,000 average income in the remaining potentially gentrifying tracts. Moreover, the almost $6,000 increase in average incomes between 1981 and 2001 recorded by gentrifying tracts is almost double the $3,000 increase recorded by the remaining CMA census tracts. There was also a $20,000 difference between the average value of dwellings in gentrifying tracts ($175,000) and the remaining potentially gentrifying tracts ($158,000) in 2001. The 2001 census shows that over a quarter of all households within potentially gentrifying tracts were classified as low-income compared to only 16 percent in the remaining census tracts. This proportion is slightly lower for the gentrifying tracts (24 percent) when compared to the remaining potentially gentrifying tracts (27 percent).
Distance from CBD
There is a statistically significant difference in the distances to downtown between potentially gentrifying tracts and all other census tracts: potentially gentrifying tracts are about 4.6km while other tracts are on average 12.5 km from the CBD. This finding is consis10t with the literature that demonstrates the importance of proximity to downtown facilities, amenities and employment opportunities as defining features of gentrifying neighbourhoods. The fact that gentrifying tracts are even closer than the potentially gentrifying tracts, 3.7 km and 4.6 km respectively, suggests that gentrification may be dispersing into neighbourhoods further away from the CBD, a trend also noted in US inner cities (Wyly and Hammel 1999) and London (Hamnett 2003) during the 1990s.
Population and Dwelling Density
The potentially gentrifying tracts have more than twice the population and dwelling densities than other tracts, while the highest densities are recorded by the gentrifying census tracts. This may be the result of high-rise development in some of the inner city census tracts. It is also a function of the extremely small average areas of potentially gentrifying tracts (less than 1sq.km.) when compared to other tracts (13.5sq.km). Changes in population and dwelling densities between 1981 and 2001 show different trends for the potentially gentrifying and other census tracts; the potentially as well as the gentrifying tracts declined in population density yet increased in dwelling density. This is due to the declining persons per households and conversions of commercial facilities to residential uses. It also suggests that gentrification contributes to the spread of the urban area. The declining population densities of Canadian inner cities were already noted by Bunting et al (2002).