By many accounts gentrification involves the loss of affordable older inner city housing through their renovation and upgrade by middle- and upper-income households. The displacement of low-income households from older housing disrupts what, in theory, should occur as a result of filtering; older inner city dwellings are made affordable by their abandonment by higher income households seeking newer dwellings at the urban periphery. Thus, the potential opportunities for gentrification and its greatest potential to displace low-income households is be found within the lower income inner city neighbourhoods with a high proportion of older dwellings.
Yet the bulk of published research on gentrification does not explicitly focus on the age of a neighborhood’s housing stock. Most authors opt for a panoptic definition of gentrification which includes processes that result in the provision of new dwellings through conversions or redevelopment as well as through the renovation and upgrade of existing housing units (Beauregard 1990, Criekingen and Decroly 2003, Ley 1988, 1993, 1996). However, this inclusive definition of gentrification does not inform the relationship between gentrification and a city’s older housing stock. This paper fills this gap by using a definition of gentrification that focuses on the renewal of the old stock, not redevelopment, and the displacement of lower-income households through reverse filtering. It examines Canadian neighbourhoods that retained a high proportion of older housing stock in 2001, but were identified as low-income neighbourhoods in 1981 and experienced income and rent increases, the socio-economic transformations associated with gentrification as noted in the literature (Beauregard 1990, Kennedy and Leonard 2001, Ley 1988, 1990).
The focus on the older lower-income neighbourhoods allows: i) a greater focus on gentrification’s impact on the filtering process, ii) an assessment of the extent of gentrification within the urban spatial structure, iii) identification of the characteristics of the older and poorer neighbourhoods that were experiencing the greatest income and rent increases, iv) an assessment of the geography of gentrification based on a stricter definition, and v) an appraisal of the neighbourhoods at risk of change, i.e., upgrade, that can be used to help target policy initiatives that compensate for the loss of affordable housing.
Attempts to define gentrification have a long, complex and disputed history (Hamnett 1991, Lees 2000, Redfern 2003). While some consensus has emerged that gentrification is a ‘chaotic concept’ (Beauregard, 1986, p. 40 & 1990) due to the diversity of factors and processes that the term covers, Criekingen and Decroly (2003, p. 2453) note that, “there is still no unanimously approved empirical delimitation of the concept of gentrification”. Hamnett (1984), however, provides a rather comprehensive definition:
Gentrification commonly involves the invasion by middle-class or higher-income groups of previously working-class neighbourhoods...and the replacement or displacement of many of the original occupants. It involves the physical renovation or rehabilitation of what was frequently a highly deteriorated housing stock and its upgrading to meet the requirements of its new owners. In the process, housing in the areas affected, both renovated and unrenovated, undergoes a significant price appreciation. (Hamnett, 1984, 284)
This definition has expanded to link gentrification with broader trends in urban renewal and social upgrading of parts of the inner city. For example, Millard-Ball (2000, 1673) defined gentrification as the “social and physical ‘upgrading’ of a residential neighbourhood”. Similarly, Ley (1993, 232) defined gentrification simply as “...an upward movement in the social status of a census tract”, while he notes elsewhere that the focus on renovation of older housing stock is an earlier and more narrow definition (Ley 1991, 330). In a more recent article, Criekingen and Decroly (2003, 2454) see gentrification as consisting of “the transformation of deprived, low-income neighbourhoods into new wealthy areas based on population change...and on improvements to the built environment”, where the authors refer to improvements to include both the rehabilitation of older buildings and the construction of new buildings on vacant land.
Bourne (1993a) divides the literature’s approach to gentrification into two main categories, a restrictive and an inclusive definition. Specifically, he notes that
...the restrictive definition refers to the invasion and succession of a neighbourhood occupied by members of one social rank or class by those of another and higher classes. Specifically, the transitions of interest are restricted to those that...involve the displacement of working class households by middle- and upper-income professional households. The second and more inclusive definition incorporates other forms of neighbourhood change, but without the necessity of residential succession or the displacement of lower-class households (emphasis in original) (Bourne, 1993a, 97).
The differences between the two sets of definitions lie in their emphasis on either rehabilitation-renovation or redevelopment. Ley (1991, 330) does note that gentrifying neighbourhoods usually exhibit examples of both renovation and redevelopment yet the link or apparent inter-relationships between these two processes has not been fully explored or conceptualized.
The stage-model of gentrification, however, suggests a link between a restrictive, i.e., renovation, and inclusive, i.e., social upgrading, definition. The process of gentrification has been presented as a stage model whereby the initial stage is set by economically marginal professionals such as artists and teachers who value the inner city locations and contribute, mainly through their sweat equity, to the upgrading of older buildings (Ley 1986). The actions of these pioneer gentrifiers may create a new and distinctive environment that attracts the attention of doctors, lawyers, engineers and the upper-level of professionals who contribute substantial amount of financial capital to the upgrade of their neighbourhoods. This initial stage of neighbourhood gentrification is usually dominated by the renovation of older dwelling units rather than redevelopment or infill development, the result of mainly individual or household-led gentrification rather than private sector institution-led gentrification (Beauregard 1986, 53). The upgrade is visible and attracts the attention of real estate agents and developers who want to speculate on the redeveloping parts of the inner city dwellings. This stage of gentrification is often associated with private sector institutional-led gentrification and involves redevelopment in-addition to renovation. A vast literature has examined the various actors and institutions involved in stages of gentrification (Bridge 2001, Ley 1991 & 2003, Redfern 2003, Smith 2002).
This study restricts the definition to a stock renewal or conversion process, not a neighbourhood redevelopment project. The definition helps us study the changes in the neighbourhoods that retained a high proportion of the older, the pre-1946 housing stock. Dwellings built before 1946 are more likely to contain the architectural features, neighbourhood attributes and proximity to downtown that are well-documented characteristics of gentrified (or gentrifying) neighbourhoods (Ley 1993, 2003, Jager 1986). Moreover, the prevalence of pre-1946 dwellings was found to be one of the best predictors of the potential for social upgrading of inner city neighbourhoods in both Canada and the United States (Ley 1993, 250; Helms 2003, Laska et al 1982). We also suggest that the older neighbourhoods represent the total supply or opportunity for renovation-led gentrification. To be sure, not all of this supply will have the conditions or the specific characteristics necessary for gentrification to take place. Yet by examining the twenty-year socio-economic transformation of older-inner city neighbourhoods, as measured by variables associated with gentrification, we can begin to assess the extent and impact of gentrification on the city’s oldest housing stock and on the filtering process.
Renovation-led gentrification, as defined above, is one way in which housing filters-up. The interest of this paper is to measure the extent and characteristics of neighbourhoods that are vulnerable to filtering-up, i.e., renovation-led gentrification. It raises a number of questions: How is a neighborhood’s “vulnerability” to gentrification measured? What makes a neighbourhood vulnerable to gentrification? What is the extent and what are the attributes of neighbourhoods that are vulnerable to gentrification?
A plethora of case study research, for example Toronto (Filion 1991, Bourne 1993a), London (Atkinson 2001, Hamnett 2003), Chicago (Helms 2003), Philadelphia (1990), New Orleans (Laska et al 1982) and Stockholm (Millard-Ball 2000), has provided a somewhat coherent portrait of the basic socio-economic and environmental attributes of gentrifying neighbourhoods (i.e.,young childless professional households, close to CBD, heritage homes etc). However, only a few studies are based on evidence from a larger sample of neighbourhoods taken from more than one metropolitan area. Specifically, Ley (1988, 1993) has provided a comprehensive examination of gentrification at census-tract level for six Canadian CMAs for two-time periods, 1971 to 1981 and 1981 to 1986. Hamnet and Wyly (1999) used discriminant analysis to classify gentrifying tracts from eight US SMSAs from 1970 and 1990, while earlier research by Lipton (1977) looked for evidence of central city revival in 20 US SMSAs during the 1960s.
Our paper assesses the trends and impacts on affordable housing stock by focusing on the socio-economic transformations indicative of gentrification within older and poor inner city neighbourhoods across 10 CMAs over the twenty-year period from 1981 to 2001. This work complements and advances the seminal work of Ley (1988, 1993, and 1996) which described the attributes and geography of gentrifying neighbourhoods during the 1970s and early 1980s. Ley’s (1988) profile of Canadian inner city tracts likely to undergo middle-class resettlement includes attributes such as proximity to elite areas, socio-economic status, and distance to various amenities, ethnicity and life cycle characteristics. His follow-up research finds that many of the key variables that had a strong correlation with his gentrification index during the 1970s no longer exhibited the same significance during the 1980s (Ley 1993). He, therefore, concluded that the geographic pattern of gentrification was more chaotic and complex during 1980s as compared to the 1970s and that “a continuing upward spiral in middle-class reinvestment” within the inner city is having regressive consequences. He warns about the “need to preserve endangered and rapidly disappearing stocks of affordable housing” (Ley 1993, 253).
Furthermore, where Ley (1988, 1993) identified gentrification in terms of a social index composing of employment and educational variables, we develop an index using principal component analysis and then screen the results using quantitative and qualitative methods: the tracts are screened using income and age of dwelling variables, and; the results are checked and further refined through key informant interviews. We also map the geographic patterns of the gentrifying tracts to test Ley’s assertion that a more chaotic spatial pattern is unfolding within the urban spatial structure.
This study develops a socio-economic profile of gentrifying tracts with some of the same variables used by Ley (1988, 1993) and then compares this profile to other tracts that appear to be potential targets for gentrification and to all the other clearly non-gentrifying tracts in each CMA. This study, therefore, builds on the typologies of inner city neighbourhoods developed by Bourne (1993), Criekingen & Decroly (2003) and Wyly & Hammel (1999). Particularly, it avoids some of the conceptual problems with respect to both the inclusive and restrictive definitions of gentrification noted by Bourne (1993, 97-98). For example, by mapping the transformations of inner city neighbourhoods, we are able to refine our search for renovation-led gentrification by omitting upgrading of established elite/high-income inner city neighbourhoods as well as sites of extensive redevelopment on grey-field sites.