Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Priced out : a profile of tenant households and their capacity to enter homeownership in metropolitan… Mendez-Gonzalez, Juan-Pablo 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-0582.pdf [ 8.05MB ]
Metadata
JSON: 831-1.0092728.json
JSON-LD: 831-1.0092728-ld.json
RDF/XML (Pretty): 831-1.0092728-rdf.xml
RDF/JSON: 831-1.0092728-rdf.json
Turtle: 831-1.0092728-turtle.txt
N-Triples: 831-1.0092728-rdf-ntriples.txt
Original Record: 831-1.0092728-source.json
Full Text
831-1.0092728-fulltext.txt
Citation
831-1.0092728.ris

Full Text

PRICED O U T : A P R O F I L E O F T E N A N T H O U S E H O L D S AND T H E I R C A P A C I T Y T O E N T E R H O M E O W N E R S H I P IN M E T R O P O L I T A N C A N A D A by JUAN-PABLO M E N D E Z - G O N Z A L E Z B.B.A. (Hon.), Simon Fraser University, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF ARTS In THE F A C U L T Y OF G R A D U A T E STUDIES (Geography) THE UNIVERSITY OF BRITISH C O L U M B I A August 2006 © Juan-Pablo Mendez-Gonzalez, 2006 A B S T R A C T A booming housing market in the early years of the twenty-first century has favoured hundreds of thousands of home-owning households in metropolitan Canada, doing much in the process to jumpstart an unwinding economy. But rapidly rising house prices have also revived the twin spectres of declining affordability and a widening economic gap between tenants and homeowners. Given the serious long-term negative consequences of tenure-driven polarisation, it becomes increasingly necessary to extend current research beyond its prevailing emphasis on the risk of downward housing mobility, and engage in complementary analyses of the ability of households to achieve an upward transition of tenure. My thesis therefore examines the financial capacity of tenant households to enter homeownership in Montreal, Toronto and Vancouver (MTV), exploring how various socio-demographic characteristics affect their likelihood to qualify for a mortgage on a moderate or average priced dwelling. M y reference point in time is 2001, a year marked by a Census of Population, a bottoming out of real house prices, falling interest rates, and the first stirrings of a housing boom that took off in full stride the following year. Through a descriptive statistical analysis of 2001 Census data, I draw a profile of the more than one million tenant households which the market had already priced out of homeownership in M T V prior to the price explosion of 2002. Geography and economic class seemed to be more important determinants of homeownership accessibility than, other household characteristics, but certain demographic subgroups were nonetheless found to exhibit polarising tendencies along the tenure line, especially in Vancouver, the most expensive housing market. Moreover, cohort effects related to age, economic u cycles, and immigrants' time of arrival appeared to interact in defining ways with other variables, particularly those related with the life-cycle model and with immigrant status. Finally, my findings suggest thai post-facto analyses of homeownership attainment, when conducted on their own, cannot be relied upon for the purposes of studying the financial accessibility of owner occupancy. Market factors relative to tenants' incomes (such as current mortgage rates, borrowing conditions, and local dwelling prices) need to be incorporated as well. iii T A B L E O F C O N T E N T S Abstract ii Table of Contents iv List of Tables v List of Figures vi Acknowledgements viii Introduction 1 A house price boom arrives 1 The effects of rising prices on housing tenure 7 Chapter 1: Unaffordable, or financially inaccessible? 14 Affordability as a problem and the problem of attributing causality 14 Difficulties in defining and operationalising the concept 20 C M H C ' s "Core Housing Need": modifying the 'rule of thumb' '. .25 C M H C ' s 'In Need/At Least Half measure 28 Why 'financial accessibility' and not 'affordability' 28 Chapter 2: An approach to estimating inter-metropolitan capacities of entry into homeownership 32 Does uneven access to homeownership matter? 36 Capacity of entry and household characteristics 39 Capacity of entry and metropolitan geography .56 Chapter 3: A profile of tenant households and their capacity to enter homeownership. 66 Discussion and concluding remarks . 107 Insights gained from analysis of the 2001 Census data 110 Looking at the boom years 114 Works Cited... 123 Appendix 130 iv LIST O F T A B L E S Table 1: Location of immigrant and non-immigrant households, 2001 (row %) 10 Table 2: Variables used in recent North American studies on homeownership (propensities 40 Table 3: Determining metropolitan 'moderate price' figures for 2001 53 Table 4: Mortgage-qualifying income thresholds, before tax, 2000 55 Table 5: Before-tax household income ranges, 2000, Groups 1, 2 and 3 55 Table 6: Labour force characteristics by metropolitan area, July 2001 61 Table 7: Tenure distribution of households by mortgage qualifying income thresholds, 2001 68 Table 8: Comparison of homeownership accessibility income thresholds and average incomes of tenant households in core housing need, 2001 69 Table 9: Qualifying tenant households, by age of Person 1 ("Primary household maintainer") 74 Table 10: Qualifying tenant households, by household composition 77 Table 11: Qualifying tenant households, by presence of children in the household 79 Table 12: Qualifying tenant households, by marital status of Person 1 81 Table 13: Qualifying tenant households, by household size 83 Table 14: Qualifying tenant households, by highest level of schooling of Person 1 85 Table 15: Qualifying tenant households, by primary language spoken at home by Person 1 87 Table 16: Qualifying tenant households, by citizenship and period of immigration of Person 1 90 Table 17: Mortgage-qualifying tenants, by sex of Person 1 and of spouse (if applicable) 93 Table 18: Qualifying tenant households, by visible minority category for Person 1 95 Table 19: Qualifying tenant households, by ethnic category for Person 1, immigrants only 97 Table 20: Qualifying tenant households, by period of immigration of Person 1 101 Table 21: Qualifying tenant households, by generational ties to immigration of Person 1 103 Table 22: Average dwelling price appreciation, in 2004 constant dollars, 2001 to 2004 105 Table 23: Total number of non-qualifying tenant households, 2001 115 Table 24: Estimated number of households most likely to suffer long-term inaccessibility to homeownership, 2001 Error! Bookmark not defined. v LIST O F FIGURES Figure 1: The Economist house-price indices, annual change, selected markets 2 Figure 2: Real Average House Prices, Metropolitan Areas (Annual, 2004 constant dollars) 4 Figure 3: Average Economic Family Market Income, Census Metropolitan Areas, 1995-2004 5 Figure 4: House-Price-To-Income Ratios, Metropolitan Areas, 1995-2004 6 Figure 5: Shannon Orr: Factors affecting affordable housing 16 Figure 6: Bunting et al.: Factors accounting for housing affordability stress and homelessness 18 Figure 7: Annual change in average house prices, selected metropolitan areas, 1996-2004 34 Figure 8: Household distribution by tenure, Canada's six major metropolitan centres 57 Figure 9: Structural dwelling types, selected CMAs, 2001 59 Figure 10: Mortgage-qualifying tenant households, by age of Person 1, 2001 131 Figure 11: Mortgage-qualifying tenant households, by household structure, 2001 132 Figure 12: Mortgage-qualifying tenant households, by presence of children in the household, 2001 133 Figure 13: Mortgage-qualifying tenant households, by marital status of Person 1, 2001 '. 134 Figure 14: Mortgage-qualifying tenant households, by household size, 2001 135 Figure 15: Mortgage-qualifying tenant households, by highest level of schooling of Person 1,2001 136 Figure 16: Qualifying tenant households, by language used at home by Person 1, 2001 : 137 Figure 17: Mortgage-qualifying tenant households, by citizenship of Person 1, 2001 138 Figure 18: Mortgage-qualifying tenants, by sex of Person 1 and of spouse (if applicable), 2001 139 Figure 19: Mortgage-qualifying tenant households, by visible minority status of Person 1, 2001 140 Figure 20: Tenant households, by ethnic group of Person 1, immigrants only, 2001 (Montreal) '. 141 Figure 21: Tenant households, by ethnic group of Person 1, immigrants only, 2001 (Toronto) 142 vi Figure 22: Tenant households, by ethnic group of Person 1, immigrants only, 2001 (Vancouver) 143 Figure 23: Mortgage-qualifying tenant households, by immigration cohort of Person 1, 2001 144 Figure 24: Qualifying tenant households, by generational ties to immigration of Person 1,2001 145 vii A C K N O W L E D G E M E N T S I wish to thank my advisors Dan Hiebert and Elvin Wyly for the support and guidance they have always so generously provided. I also wish to acknowledge some of the people who in various ways have made this thesis possible: Fiona Jeffries, Jennifer Hyndman, Juanita Sundberg, Zoe Druick, David Ley, Kevin Gould, Tom Durning, Jose Aparicio, Junnie Cheung, Ryan Lucy, and the staff, faculty, and graduate students in the Department of Geography. Finally, I also extend a very special thanks to my parents, Julio and Marina. viii INTRODUCTION A house price boom arrives It took only three or four of the new millennium's first six years for North America's metropolitan house prices to reach new historical highs. This rapid appreciation of urban real estate markets—some of which were still recovering from earlier meltdowns at the outset of the boom—is especially remarkable considering the sore state of economic affairs that welcomed the twenty-first century. The crash of the N A S D A Q index in March of 2000 had unceremoniously punctured the hi-tech bubble of the 1990s, spreading uncertainty and panic across the world's stock markets. By the summer of 2001, enormous amounts of capital were threatened by volatile share prices, dropping profits, and sluggish industrial growth. Fearing massive job losses, cities began to brace for the worst. For investors, the almost reflexive response was a hasty search for the next spatial fix, which quickly led to residential real estate and such corollary industries as finance and construction. Two crucial ingredients came together at that point to yield the remarkable degree of capital switching that ensued: first, a growing awareness of demographic urban pressures, and second, a series of moves by the US Federal Reserve and central banks elsewhere to lower the price of credit. Lower interest rates not only made housing markets more attractive to investors, but also pulled in hundreds of thousands of regular folks eager to 'trade up' their owner occupied dwelling, or in the case of tenants, to purchase a first home. By 2002, frenetic market activity had triggered 1 Figure 1: The Economist house-price indices, annual change, selected markets (Based on nationwide average nominal house prices) 02001 (Jan-Dec avg.) • 2004 (July- Sept avg.) • 2005 (July Sept avg.) Note: Figures for Italy are provisional. Adapted from: The Economist, March 28, 2002 & December 8, 2005. a vertiginous rise in dwelling prices and mortgage lending and refinancing, not just in North America, but also in Europe and Australia (Figure 1). As a result, national economies surged despite low rates of productive investment and real household income growth, fuelled by homeowners who, prompt (and aggressively prompted) to cash in on this boom, began to inject enormous amounts of liquidity into consumer markets. It was as i f the "irrational exuberance" that characterised the dot-com heyday had found a new medium in the materially grounded world of residential real estate. Since then, there has been no dearth of concern over the pace of such gravity-defying house price growth. As early as March 2002, for example, the influential magazine The Economist spoke of "mounting evidence of house-price bubbles in some big cities," where house-price-to-income ratios were "perilously close" to their historical peaks (Economist, 2002). But not even the frequent uttering of the word "bubble" would curtail the impetus to build, renovate, sell high, and (not for long) buy low. For scores of 2 investors throughout much of the Global North, the accelerated revaluation of housing stock and urban land has been too impressive to simply stay at the margins (Figure 1). Moreover, prominent players in the global housing industry have opposed any characterisation of the situation as exuberant or volatile. Among them is Canada's national housing agency, the Canada Mortgage and Housing Corporation (CMHC), which recently stated that there is "No housing bubble" in Canada, and that the "upward pressure on prices [is] supported by solid fundamentals" (CMHC, 2005: 36). The case of Canada may or may not be exceptional, but in any event it is one that is worth examining. The rise of residential prices has resulted in a remarkable transformation of the residential landscape, especially in the largest metropolitan areas. As a result of adjustments associated with the aggressive incorporation of Canada's two largest metropolitan real estate markets into the global circuits of capital in the 1990s (Ley and Tutchener, 2001; Olds, 2001), price gains prior to 2001 had been characterised by their unevenness (Figure 2). While in real terms the average price in Toronto had been rising since 1996, it was practically stagnant in Montreal during the second half of the 1990s. In the meantime, the price drop that followed the boom of the first half of the decade in Vancouver continued unabated until 2001. By 2004, however, the situation had changed dramatically. Inflation-adjusted average market prices in these three C M A s reached highs unseen in more than 20 years. While the 2004 real average price in Vancouver was only 1.6 percent higher than in 1995 ($373,900 versus $368,000 nine years earlier), it is nonetheless striking to see real average prices come out of a six-year fall in just three. Meanwhile, prices have 3 Figure 2: Real Average House Prices, Metropolitan Areas (Annual, 2004 constant dollars) -Vancouve r "Toronto - Montreal 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Note: Average yearly sales prices as per the Canadian Real Estate Association's Multiple Listings Services (MLS) for each Metropolitan area, reported by the C M H C and adjusted by the author using Canadian average Consumer Price Index figures. C R E A ' s definition of Metropolitan Area differs from Statistics Canada's definition of CMAs. Sources: CMHC/Canadian Real Estate Association; Statistics Canada. continued to soar; the Real Estate Board of Greater Vancouver (http://www. realty•link. org/statistics/buyersjnarket_HPIjnain.cfm), for example, estimates that in May of 2006, the benchmark sale price of the "typical" detached dwelling in this city-region was $635,926, or 89.4 percent higher than it had been only five years earlier—when Vancouver's real average price hit the bottom of the trough. Intuitively, one would expect soaring house prices to reflect fast growing income levels. But in Canada's largest metropolitan centres, that picture is perplexingly distorted. Since the turn of the new millennium, the rate of growth in yearly average incomes has been modest at best (Figure 3), introducing a likely source of compounding anxiety for those who remain uneasy about the possibility that a price bubble has enveloped many of Canada's residential real estate markets. In 2004, inflation-adjusted average family 4 Figure 3: Average Economic Family Market Income, Census Metropolitan Areas, 1995-2004 (Annual, 2004 constant dollars) "•Vancouver —-Toronto - Montreal 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Note: Includes economic families and unattached individuals. A n economic family is defined as a group of two or more persons who live in the same dwelling and are related to each other by blood, marriage, common law or adoption. A n unattached individual is a person living either alone or with others to whom he or she is unrelated, such as roommates or a lodger. Sources: Survey of Consumer Finances; Survey of Labour and Income Dynamics. income (before tax) in Vancouver was, much like the average house price, only slightly higher than in 1995, allowing Montreal to almost catch up to its range in the mid-$50,000-per-year. But Montreal's impressive income growth over the previous eight years was still insufficient to lift the C M A ' s average above Vancouver or Toronto. In fact, income growth was uneven across the three CMAs throughout the 10-year period between 1995 and 2004. After the gains of the late 1990s, real average family income in Montreal and Vancouver grew almost imperceptibly between 2001 and 2004, a period that saw house prices in both cities post steep gains. Only in Toronto did high average pre-tax family incomes seem to keep pace with the speed at which average house prices grew after 2001. Averages are obviously sensitive to outliers at the top, and one can indeed expect Canada's main global city to be 5 home to a fair number of high earners. Nonetheless, the average income-to-average-house-price relationship after 2001 is consistent with a pattern that dates at least as far back as 1995. But a look at house-price-to-income ratios confirms that in Canada's two other largest cities, the decoupling of dwelling prices from incomes grew larger during the period between 2001 and 2004, and remained at a relatively high level in Toronto (Figure 4). While Montreal's ratio inched closer to Toronto's, in Vancouver it surged back to the high levels of the mid-1990s. As a result, the average house price here was equivalent to almost seven times the annual average market income of economic families in the city-region by 2004 (a ratio that was nonetheless lower than it had been at the peak in 1995). In contrast, Demographia International (2006) recently reported that the median house price in Edmonton, one of Canada's ten fastest growing cities, was 2.4 times the Figure 4: House-Price-To-Income Ratios, Metropolitan Areas, 1995-2004 (Calculated from prices and incomes in 2004 constant dollars) ^-Vancouver ~~ Toronto - - Montreal 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Note: Based on consumer inflation-adjusted average yearly sales prices, as per the Canadian Real Estate Association's Multiple Listings Services (MLS) for each Metropolitan area, reported by the C M H C . C R E A ' s definition of Metropolitan Area differs from Statistics Canada's definition of CMAs. Income refers to the yearly average economic-family income for each C M A , as reported by Statistics Canada in the Survey of Consumer Finances and the Survey of Labour and Income Dynamics. Sources: CMHC/Canadian Real Estate Association; Statistics Canada. 3.00 4 6 city's median household income in late 2005. While this median figure is not strictly comparable with the averages I have reported here, it does provides some indication of the enormous gulf between incomes and house prices in Montreal, Toronto and Vancouver relative to the situation of other metropolitan areas in Canada. The effects of rising prices on housing tenure At the beginning of the price boom in 2001, five of Canada's six largest metropolitan centres recorded homeownership rates in the 60 percent range or higher. (Montreal was the exception, with a rate of just under 50 percent.) For these hundreds of thousands of home-owning households, booming dwelling prices should normally translate into an improved equity position. And indeed, many have chosen to materialise net asset gains derived from the explosive appreciation of their property by either selling it, borrowing new money against this accrued equity, or refinancing their mortgages at lower interest rates than in the 1980s and early 1990s. The resulting cash windfalls are often spent on consumer goods, but also on home renovations, second homes, or trading up to a larger and better-located dwelling (Engeland et al., 2005). Low interest rates have also encouraged many tenant households to leave the rental market and purchase a first home (CMHC, 2005). The real estate industry could not be happier. According to one of the largest realtors in Canada (ReMax, 2005), annual sales volumes in 2005 were more than double those of ten years earlier in cities like Ottawa, Calgary, Toronto, Montreal, and Edmonton. Unfortunately, this housing boom has not benefited everyone equally. As I will show in this thesis, expensive housing markets in metropolitan Canada have turned homeownership into a distant, perhaps impossible dream for hundreds of thousands of 7 households with insufficient incomes. While owner occupancy is not the only mode of tenure and its absence does not generally equate with homelessness, a housing market that shuts out large numbers of potential buyers—particularly those at key stages of their life cycle (Boehm and Schlottmann, 2002)—can harbour a host of other socio-economic and environmental impacts, including the enhancement of polarisation along class and racial lines (Boehm and Schlottmann, 1999; Chinloy, 1999; Gyourko et al., 1999; Hackworth and Wyly, 2003; Hulchanski, 2004). For this reason, it is important to assess the capacity of households to enter homeownership, particularly at times when access to this form of tenure may appear to be overly constrained by income levels that fail to keep up with rising house prices. Situating my research in the year 2001, my thesis seeks to portray the situation of the financial accessibility of homeownership in metropolitan Canada on the eve of the new millennium's house-price explosion. To this end, I will analyse data from a variety of sources (including the 2001 Census of Canada) using the household as unit of analysis. My analytical approach will be the statistical profile, focusing in particular on tenant households that were more likely to experience homeownership as nothing more than a lofty aspiration, already beyond their economic means in the year before the price rally took hold of metropolitan housing markets in Canada. This thesis also devotes considerable attention to the spatially, demographically and socio-economically variegated character of my subject population. As I will show, the likelihood that a tenant household is financially able to enter owner-occupancy varies across metropolitan areas of residence. Inter-metropolitan unevenness is also manifest in the distribution of financial accessibility among households with different socio-8 economic and demographic characteristics. Of course, there is also a great deal of variation in the relationships between dwelling prices and socio-demographic characteristics within metropolitan areas, but an analysis of multiple metropolitan areas at the census tract or census sub-division scale would be beyond the scope of this thesis. My research will examine select metropolitan areas comparatively, focusing on household structure but also on key characteristics of the Primary Household Maintainer (i.e., the first person identified by Census respondents as being responsible for household payments, normally the person who contributes the greatest amount toward these expenses), including sex, age, education level, visible minority status, ethnicity, immigrant status, and generational ties to immigration. To keep this descriptive statistical analysis to a manageable size, I will only examine Canada's three largest metropolitan areas: Toronto, Vancouver and Montreal—three very different housing markets that together accounted for fully one third of all households in the nation at the time of the 2001 Census. These three cities were also the main immigrant destinations in Canada, taking in 59 percent of all immigrant households and 73 percent of those with five years or less in the country (Table 1). Immigrants account for a growing proportion of Canada's population, a situation that is likely continue for years to come (see Hiebert, 2005). Not surprisingly, their housing careers have been the subject of considerable research in recent years. We have learned for example that immigrant groups with stays of 20 years or more in Canada tend to have homeownership rates above those of non-immigrant households, but we also know that recent immigrants tend to fare poorly in the housing 9 Table 1: Location of immigrant and non-immigrant households, 2001 (row %) Canada (Number) Montreal CMA Toronto CMA . Vancouver CMA Rest of Canada All households 10,805,600 12.4 14.3 6.5 66.7 Non-immigrant* 8,439,400 12.5 8.6 5.0 74.0 Immigrant* 2,319,900 12.1 35.0 12.1 40.8 pre-1976 1,222,400 10.8 30.1 10.0 49.1 1976-1985 358,900 13.7 34.9 12.7 38.6 1986-1990 243,700 13.0 43.7 13.0 30.3 1991-1995 269,800 13.2 43.3 16.8 26.7 1996-2001 225,100 14.2 42.0 16.5 27.4 * Refers to the immigration status of the Primary Household Maintainer (PHM). Households where the P H M is a non-permanent resident are not included in this category. Adapted from C M H C (2004). market, particularly during the first ten years of settlement (Engeland et al, 2005; Hiebert, 1999; Laryea, 1999; Ley and Smith, 1997, 2000; Mendez et al, 2006). Little is known, however, about the way in which demographic variability and the socio-economic conditions of newcomers in each of Canada's main destination cities affect the . ability of immigrant households to access owner occupied housing. This thesis can in part be seen as step towards addressing this research gap. My motivations are not simply academic, though. Developing a more sophisticated understanding of the housing affordability situation in Canada is an important project in its own right. Regardless of the immigration status and tenure of affected households, affordability problems can have a variety of negative impacts on the social and economic well being of individuals, including excessive household debt and depletion of savings, personal hardship, relationship breakdowns due to stress, health issues related to stress or to households resorting to overcrowding, and an increased risk of homelessness (Gabriel et al, 2005), while for immigrants in particular it may hinder the process of adaptation to a new country (Ley et al, 2001; Murdie and Teixeira, 2003). 10 Equally important is the fact that individuals do not bear the costs of a lack of affordability alone; depending on its intensity and duration, this problem can have serious economic, social, spatial and environmental consequences for governments, the business sector, and society in general. In a recent review of the literature, Gabriel et al. (2005) identify among these wider negative effects the diversion of capital away from productive investment, wage pressures, heightened population mobility, political tension, socio-economic polarisation, spatial concentration of low income households in low-rent areas, and disincentives to develop sustainable building techniques and environmentally friendly urban services. But the costs and risks associated with affordability problems are hard to measure, and causality is difficult to isolate, attribute and understand. Moreover, the concept of affordability itself is a difficult one to define. In Chapter 1,1 will examine the challenge of identifying determinant factors behind affordability problems, and discuss the difficulties of developing a generally accepted definition of affordability. This discussion will include an overview of the concomitant difficulties that researchers face in trying to operationalise the concept's various meanings empirically. The chapter therefore provides the groundwork for Chapter 2, which begins by considering the relevance of studying the accessibility of homeownership, and then provides an overview of my research methodology. Chapter 3 contains the results and analysis of my investigations, including a focus on various characteristics of immigrant and non-immigrant households considered to have insufficient incomes for accessing homeownership in Montreal, Toronto and Vancouver. I conclude this thesis with an overview of my findings and some suggestions for further research. 11 As I will show in this thesis, thousands of metropolitan tenant households would have been able to access owner occupancy in 2001 by opting for modestly priced dwellings. Indeed, 2001 was the best possible time to purchase a house in the ten-year period from 1995 to 2004, due to historically low mortgage rates and the lowest average income to average house price ratio in years (Figure 2). But more importantly—both numerically and in terms of the social meaning and consequences of this sobering fact— there were hundreds of thousands of households that in 2001 were already financially shut out of ownership, even at the peak of the buyers' market and before the price boom hit its stride in 2002. This inequality of access, suffused with its own inter-metropolitan geography, matters increasingly when housing prices soar. As I will discuss in Chapter 2, the ability to accumulate wealth through homeownership has in recent years resurfaced as a key driver of widening socio-economic polarisation. For certain types of household, being unable to access owner occupancy increases the likelihood of remaining at, or falling to, the bottom end of the socio-economic measuring stick. Each section of my thesis seeks to place this fact in a theoretical, methodological, and empirical context, concentrating on issues of financial accessibility of homeownership at the turn of the new millennium in Montreal, Toronto and Vancouver. Taken together, these chapters should provide a novel portrait of those tenant households that seemed more likely to be priced out of the three largest housing markets in Canada at the turn of the twenty-first century. On a broader level, my thesis also seeks to fill a gap in current research on housing affordability in Canada, which in recent years has placed an emphasis on the risk of downward mobility in tenure. I am referring to the preference given by housing researchers to shelter cost-to-income ratios and the C M H C ' s 12 concept of Core Housing Need, two measures that implicitly highlight the risk faced by certain households of becoming homeless or of losing their mortgaged dwelling due to excessive spending on housing relative to income. My thesis concentrates instead on the availability of financial possibilities for households to transition upwards into homeownership. In this sense, I hope that a more complete picture of the housing affordability situation in metropolitan Canada will emerge. 13 C H A P T E R 1: U N A F F O R D A B L E , OR F I N A N C I A L L Y INACCESSIBLE? Affordability as a problem and the problem of attributing causality The social, economic, and political impacts that are generally associated with housing affordability problems constitute a type of phenomenon that can hardly be attributed to one single cause. The costs and risks of a lack of affordability, including the spatial concentration of poverty and growing socio-economic polarisation (Boehm and Schlottmann, 1999; Chinloy, 1999; Gyourko et al, 1999; Hackworth and Wyly, 2003; Hulchanski, 2004), are not only linked to this one issue, but may also be associated with a wide combination of trends, processes, structural arrangements, and public or private decisions that stretch beyond matters of household incomes or expenditures and house prices and rents. Discussions that focus on simple cause-effect relationships are bound to prove inaccurate and unreliable. Despite this avowed difficulty, the academic and policy literature is not short on interpretations of the ways in which housing affordability problems develop. (This thesis not an exception: in the first page of the introduction, I briefly sought to situate the origins of the recent house price boom in Canada within the context of the first over-accumulation crisis of the twenty-first century.) Caution is needed when analysing this literature. Gabriel et al, 2005 point out that research often isolates or emphasises factors that are likely to affect the interests of the agency or lobby that sponsors the study, which inevitably establishes a linkage between the attribution of causality on the one hand and recommendations for sector-specific solutions on the other. Examples of this pattern include: a discussion paper on housing affordability in Canada published by TD Bank Financial Group (TD Economics, 2003), in which capital taxes are singled out as one of 14 the main causes of the deficit of affordable rental housing stock; a speech by Greg Suttor (2004) of the City of Toronto, pointing an accusatory finger towards the inadequacy of federal (and to a lesser extent provincial) government funding for ongoing operating expenditures associated with the provision of affordable housing (such as rent supplements for tenants currently housed in the private rental sector); and the Toronto Board of Trade's (2003) study on housing affordability, which concludes that the federal government's efforts "to close loopholes in the federal tax system have discouraged private investment in rental housing" (11). A research focus on sector-specific factors does not necessarily discount the validity of a study's conclusions, and my goal in citing these examples is not to engage in a discussion on the merits of their recommendations. Instead, I simply want to point out that such a circumscribed research approach is more likely to miss other equally important conceptualisations and explanations of the problem. In contrast to these sector-centric approaches, Shannon Orr (2000) and Bunting et al. (2004) have each proposed more systematic and comprehensive models for describing the sources of current housing affordability problems in Canada. For Orr (2000), housing affordability has two components: the cost of housing and the ability to pay for it. Each of these elements, she explains, is affected by various factors such as urban growth and the supply of dwelling units for the former, and household incomes and mortgage rates for the latter (Figure 5). The interplay of these variables will have an impact on the ability of households to secure "some standard of housing (or different standards) at a price or a rent which does not impose, in the eyes of some third party (usually government) an 15 Figure 5: Shannon Orr: Factors affecting affordable housing Energy costs Supply of housing Taxes/levies Market-based assessment Urban growth Household incomes Mortgage rates Household incomes Mortgage rates Welfare/employment/ pension reform Adapted from Orr (2000: 6; 13; 15) unreasonable burden on household incomes." (Maclennan and Williams, 1990, cited in Orr, 2000: 2). Readers with keen geographical imaginations will notice that time and space are two obvious gaps in Orr's model. As discussed earlier, different cities (and different neighbourhoods within each city) can be more or less affordable than others at different points in time. Sensitivity to these differences is an underlying feature of Bunting et al.'s (2004) model. The authors schematise their own understanding of the factors that affect what they call "housing affordability stress," meaning the subjectively experienced 16 Cost of housing Ability to pay AFFORDABLE HOUSING adverse relationship between income and housing costs (Figure 6). The-phenomenon, they argue, is the product of deep changes over time in the economic, social/demographic, and policy contexts at various geo-jurisdictional scales. Each of these major shifts has had its own chain of general and place-specific consequences, which in turn have led to a) an increase in the numbers of low-income households, b) a rise in rents and house prices, and c) a decrease in the supply of affordable housing. Bunting et al. (2004) argue that these last three factors together have unevenly increased housing affordability stress levels within and between Canada's metropolitan areas. Ultimately, both Orr's (2000) and Bunting et al.'s (2004) schemas attribute Canada's affordability problem to a combination of declining incomes and the inadequacy of the housing supply. It is worth noting, though, that not everyone agrees with this dual causality model. Moore and Skaburskis (2004), for example, show that Ontario's affordability problem—expressed in terms of the proportion of the population spending more than 30 percent of their household income on rent—worsened between 1991 and 1996, a period during which provincial rent controls were in place. To these authors (2004: 406-7), the incidence of declining affordability during a time when such policies were in place suggests that: "the growth of affordability problems is fundamentally related to [downward] changes in income rather than the inadequacy of housing supply." In turn, the causes of the decline in average household incomes are the source of further disagreement. As Bunting et al. (2004: 367) observe, academic researchers and service providers tend to differ with conservative commentators, who often blame "'individual' factors such as disability or substance abuse [instead of] 17 Figure 6: Bunting et al: Factors accounting for housing affordability stress and homelessness ECONOMIC CHANGE Polarised, post-industrial occupational profile Uneven metropolitan growth rates Overheated real estate in fast growing CMAs SOCIAL/DEMOGRAPHIC , ^ CHANGE ^ Demise of family supportive wage More dual-earner households (driving up demand for larger and/or better quality housing) More single-person households More single-parent, female headed households More immigrants Increased numbers of low-income households Increased house prices and rents Increased housing affordability stress POLICY CHANGE Cut-back in social support Elimination of rent controls Private production of rental housing negative affected by locally generated land-use policy and of municipal and federal taxation changes Demise of social housing Increased house prices and rents Decreased supply of affordable housing Adapted from Bunting, Walks and Filion (2004: 365-6) 18 'stractural' changes in contemporary society." It is therefore critical not to ignore the political and ideological character of the housing affordability question. Surprisingly, Bunting et al.'s (2004) schematic representation of the social/demographic factors accounting for housing affordability stress does not mention an aging population, perhaps the single most important demographic challenge faced by Canada today. Other factors that have been associated with issues of affordability but are not included in Orr's (2000) and Bunting et al.'s (2004) models include the cost and availability of financing—not just interest rates but also the regulatory changes and industry innovations that have increased the supply of mortgage funds since the post-war years (Carter, 1997) —and the flourishing informal housing market (including basement suites in duplexes and single family dwellings), to which growing numbers of tenants have turned partly in response to the insufficient supply of purpose-built rental stock (Engeland et al, 2005). Additionally, there are certain critical place-effects to consider. Moore and Skaburskis (2004:396) have suggested that an increase in the size of the wealthy population "can inflate land prices and raise dwelling rent" and cite Quigley et al.'s (2001) analysis of San Diego and San Francisco in California, where better-off households were found to exert pressure on the supply and price of housing in areas of pronounced income inequality. A final set of variables that the literature identifies as potentially affecting dwelling prices (and therefore also the incidence of housing affordability stress) include: changing consumer preferences and expectations with regards to housing; a tight regulatory context for land-use; declining rental yields for landlords and developers; rising construction costs; and a tax environment that encourages homeownership. Zeroing 19 in on this last factor, David Hulchanski (2005) notes that Canada's housing policy is primarily geared towards ownership, citing as evidence the provision of tax shelters through RRSP loans for home purchases and of subsidised mortgage insurance through C M H C for households that meet the lending requirements of financial institutions. "The social and political system clearly privileges ownership housing and house owners," he argues. "There is no tenure neutrality, nor do public policies relating to housing demonstrate a progressive orientation (that is, they do not focus on those in need first before helping those who are already well-off). To call this an affordability problem among the poor is to conceal an entrenched bias in public policy." (2005: 6) Difficulties in defining and operationalising the concept Hulchanski's comment not only constitutes a critique of Canada's housing policy tools, but also reveals the instability of the idiom 'affordability problem.' Until now, I have used the terms 'affordability,' 'affordability stress,' 'ability to pay,' and 'financial accessibility' interchangeably, reflecting the often-unproblematic way in which they are commonly used. Yet, as Hulchanski's (2005: 6) above-cited comment makes plain, the precise definition of what these terms mean is anything but stable, and has been the subject of considerable debate within academia and the field of policy-making. Quigley and Raphael (2004: 191-2) explain that the word affordability ... jumbles together in a single term a number of disparate issues: the distribution of housing prices, the distribution of housing quality, the distribution of income, the ability of households to borrow, public policies affecting housing markets, conditions affecting the supply of new or refurbished housing, and the choices that people make about how much housing to consume relative to other goods. This mixture of issues raises difficulties in interpreting even basic facts about housing affordability. 20 i This lack of a fixed referent also makes affordability a difficult situation to measure empirically, due to the impossibility of capturing some of the less tangible and more subjective issues with which the concept is implicitly associated. For example, housing affordability can be understood as the continuing costs of a mortgage or rents relative to income, problems of accessing affordable housing (e.g. first home ownership), not being able to afford housing costs after meeting other expenditures, or a problem of too low an income or two high housing prices. Even more problematically, affordability can be experienced by different household types in different ways, that is, the employment, transport, health, and other consumption trade-offs that have to be made by singles, sole parents and couples with children as they adapt their circumstances to high housing costs and/or low income. ... No measure or indicator of affordability or suite of indicators can capture the nuances of how households and individuals adapt their lives to minimise or mitigate affordability problems. (Gabriel et al, 2005: 37.) To put it differently, it is unrealistic to expect any single statistic to be able to describe the effects or relationships between so many different types of factors. Despite these limitations, landlords and financial institutions typically use the shelter cost-to-income ratio as a 'rule of thumb' to make decisions over the allocation of private rental units or mortgage financing in the marketplace. According to Chaplin and Freeman (1999), the shelter-to-income ratio is the most widely used indicator of affordability, even though it has been widely criticised for its lack of clear rationale in defining a benchmark, its application of a single measure to all locations, household compositions and forms of tenure, its lack of accounting for housing quality and the incidence of overcrowding, and its incomplete assessment of the household's actual economic situation (Gabriel et al, 2005). How can such a flawed indicator be considered a bona fide norm of what society should consider as 'affordable?' 21 A brief history of the housing expenditure-to-income ratio One way to answer this question is to look at the origins of the shelter cost-to-income ratio, and Hulchanski (1995) conveniently provides a brief overview of its history. The author takes as a starting point the mid-nineteenth century, when early modern social science became convinced that social behaviour was governed by scientific laws. As part of this much larger movement, academics and government officials in Europe and North America undertook the project of formulating laws of consumption that would explain the rapport between household budgets to expenditures. For decades, many researchers centred their efforts on housing consumption, exploring in particular the income-to-rent relationship. By the late 1930s, however, these pursuits had been largely abandoned, buried under the heavy weight of numerous conceptual, theoretical, and practical problems such as the lack of standardized definitions of income and housing costs, the inadequacy of available data, and the limitations of existing techniques of statistical analysis. But i f the academic realm came to recognise as failures these original attempts to establish the laws of housing expenditure, the housing industry uncritically embraced one of them as a common-sense reflection of a social truth. This was the shelter cost-to-income 'rule of thumb.' Citing George F. Kengott's (1912) influential research, Hulchanski (1995) traces the roots of this widely used indicator to the nineteenth century maxim "one week's wage for one month's rent." At the time of his writing, Kengott argued that the adage continued to be generally apt. Having compared archival records of various late nineteenth century US cities to his sociological observations of early 22 twentieth century Lowell (a town in Massachusetts), he reported that workers in a variety of East Coast towns had for decades devoted about 20 percent of their wages to rent, light and fuel, and continued to do so in Lowell. Kengott's finding was enormously influential. The 2-to-l 0 ratio quickly became the generally accepted indicator of housing affordability, particularly in the mortgage-lending sector. By the 1950s, the 20 percent rule had given way in Canada to 25 percent, only to be replaced again by a 3-to-10 ratio (or 30 percent) in the 1980s (Bacher 1993, Hulchanski 1995). Problems with the 'rule of thumb' approach Perhaps unsurprisingly (considering its quaint but methodologically flawed origins), the 30-percent-of-income-on-housing 'rule of thumb' raises a number of technical and ethical questions that have been the object of debate for many years. From a methodological perspective, Hulchanski (1995, 2005) notes that market income is a limited measure of the ability to pay for housing, as it is only one of the many means that households have at their disposal to help them meet their long-term and day-to-day needs. There are other cash and non-cash resources to which a household may have access, and these are the result of decisions and interactions occurring both inside the home and outside. They include the extended family, close acquaintances, the neighbourhood, community-based groups and organizations, and government agencies. The extent to which households are able to tap into these formal and informal economies inevitably affects the proportion of market income that can be reasonably devoted to shelter costs, but not always in an easily measurable way. With respect to ethical concerns raised by the use of ratio measures, Hulchanski (1995) notes that as a statistical indicator, the shelter cost-to-income ratio assesses 23 households in terms of an aggregate characteristic ('low income relative to shelter costs') instead of individual ones (such as the household's ability to tap into cash and non-cash resources). When market supply decisions are based on ratio indicators, he argues, landlords and lenders act in a discriminatory fashion (because, he explains, the ratio is based on "using the category lower than average household income as a negative stereotype"), while also ignoring consumers' freedom to choose how to allocate their total household resources. More broadly, Hulchanski (1995) observes that this rule of thumb marries generalized observations of what the average household tends to spend on housing with practical assumptions about what it ought to. Its normative use as a measure of affordability is therefore misleading, and in some cases unreliable or simply invalid. He reaches this conclusion through an analysis of the most common uses of the shelter cost-to-income ratio. He argues that its deployment is appropriate when it works simply as a framework for generating analytical data without claiming for itself any 'rule of thumb' status regarding affordability or ability to pay. Examples of such uses include descriptions of household expenditures, analysis of trends and dynamics regarding the relative position of various households, and deciding on maximum income criteria for accessing public sector subsidies. But when the task at hand is to define housing need, predict a household's ability to pay rent or a mortgage, or setting up criteria to select or approve tenants or mortgage borrowers, the shelter cost-to-income ratio is deemed to be inappropriate; Hulchanski (1995) argues that in these three last cases, the function of measurement attributed to the ratio is invalidated by inadequate definitions of what counts as income, by the fact that it does not account for the degree of stability of the 24 household's income flows, and by problems associated with aggregation that are inherent to the normative use of a'rule of thumb.' C M H C ' s "Core Housing Need": modifying the 'rule of thumb' Despite long-standing criticisms such as those forwarded by Hulchanski, the shelter-cost-to-income ratio continues to be the most commonly used measure of affordability, i f only because it is simple to use, it can be easily explained to consumers and other users who are non-experts, and it "depends on a few variables that are readily available over time." (Gabriel et al, 2005: 24). In an attempt to minimise its most frequently cited limitations, a series of modifications to the basic ratio have been developed over time. These modifications alter the calculation of the ratio through the introduction of new or adjusted variables, in particular those judged capable of dealing with the lack of sensitivity to geographical, household, and dwelling characteristics (Gabriel et al, 2005). Such variables include local house prices or rents, household composition, forms of tenure, housing quality, overcrowding conditions, and alternative measures of a household's economic situation such as Statistics Canada's Low-Income Cut-Offs. Another possible modification involves not only a change in the calculation procedures that the researcher adopts, but also a semantic shift based on abandoning the term 'affordability' for 'need.' Feeman et al. (2002: 102; cited in Gabriel et al, 2005: 7) identify three conceptual differences that distinguish these two words: [First, affordability usually] looks only at housing expenditure and incomes, not housing standards, while need looks at standards and does not directly mention expenditure or income. Second, with need, the policy emphasis is on production and allocation, while with affordability the 25 emphasis is on incomes and prices.... Third, need is defined in absolute terms, while generally affordability is not.... In practice, these extremes are modified so that need tends to be measured in terms of those unable to afford social housing, while affordability accepts the standards implicit in mortgage and social sector allocation as well as unfitness and other regulations. In Canada, this practical modification of the two 'extremes' implied by the terms 'affordability' and 'need' forms the basis of C M H C ' s modified shelter cost-to-income ratio. It is called 'Core Housing Need,' and it is produced with custom Census data obtained from Statistics Canada, allowing the measure to be sensitive to household composition, over-crowding conditions, dwelling quality, and geography. The 'Core Housing Need' measure is designed to take into consideration three standards of acceptability: adequacy (housing in good condition of repair), suitability (housing that meets the needs of the household in terms of number of rooms relative to household size and structure), and affordability (defined as housing that costs less than 30 percent of pre-tax household income). As (CMHC 2004: 3) explains, a household "is said to be in core housing need i f its housing falls below at least one of the adequacy, suitability, or affordability standards and it would have to spend 30% or more of its before-tax income to pay the median rent of alternative local housing that is acceptable (meets all the three standards)." C M H C , Statistics Canada, and other public and private sector users employ this statistic to determine the proportion of tenant and home-owning households statistically judged to be experiencing need as a result of a lack of affordability. The measure can also help to identify the types of households most likely to find themselves in such a situation. The geographically sensitive standards of this modified cost-to-income ratio are implicit in CMHC's incorporation of local costs of acceptable housing into the 26 calculation of the ratio. This is one of the strengths of the Core Housing Need measure, but the way in which this objective is achieved also brings into light one of the weaknesses of this approach. Critics may point out that in trying to account for a larger degree of complexity, this ratio modification introduces new unverified assumptions into the final statistic (Gabriel et al, 2005). For example, the application of the three standards of acceptability in the calculation of core housing need assumes that alternative adequate housing is available in that specific metropolitan market but is simply unaffordable to households ranked below the acceptability benchmark. The measure, however, does not provide any indication of the validity of such an assumption in a given city or local context. Nonetheless, the C M H C ' s measure seems to satisfy many of the concerns of students of past government practices, who favour policy solutions based on a dual strategy of demand-side and supply side responses (see for example Carter, 1997; Katz et al, 2003; Pomeroy, 2004). In as much as Core Housing Need takes for granted the existence of built and vacant units in the local housing supply, it directs the user's attention to any inadequacies in the supply of affordable housing in particular: households said to be ' in core need' cannot access adequate housing in regular market conditions, and therefore require non-existing units under more favourable terms. Consequently, the focus on 'need' also implies a concern with the insufficiency of economic means for a given segment of a population, a social problem that gains expression through inadequate housing consumption. Hence, the measure succinctly evokes both demand and supply concerns within a local context. 27 C M H C ' s 'In Need/At Least Half measure Along with its Core Housing Need measure, C M H C continues to publish simple shelter cost-to-income statistics in order to provide a multifaceted portrait of affordability in Canada. In addition, it provides figures on>the proportion of households considered to be in core housing need and spending more than 50 percent of their before tax income on housing (see for example C M H C , 2004). This hybrid indicator, which constitutes a further modification of the simple shelter cost-to-income ratio, is referred to as I N A L H — an acronym for In Need/At Least Half, which is itself an abbreviation of 'households in core housing need spending at least half of their income on housing.' (Curiously, C M H C (2004: 4) notes that farm households and households reporting zero or negative incomes and those with shelter cost-to-income ratios of 100% or more are not included in the calculations of households in core housing need, as these cases are deemed to be "uninterpretable"; "Of the 11.6 million households identified in the 2001 Census, 10.8 million were non-farm, non-reserve households with interpretable shelter cost-to-income ratios (STIRs).") INALH statistics are used as a tool to determine the share of households that are deemed to be at high risk of homelessness due both to the disproportionate amount of income they devote to housing and to the lack of better priced acceptable housing locally. But risk is not the same as. incidence, and the I N A L H measure is obviously unable to predict how many of these (or other) households do in fact slip into homelessness eventually. Why 'financial accessibility' and not 'affordability' A review of statistical measures that can reliably predict which households will fall into homelessness is beyond the scope of this thesis. Nonetheless, it is worth dwelling 28 for a moment on the risk assessment aspect of the Core Housing Need and INALH measures employed by C M H C and Statistics Canada. In my reading; the risk of downward mobility of tenure is clearly the analytical priority of these two approaches to the assessment of affordability problems, both in the case of tenants (losing their social or private rental housing and becoming homeless) and of homeowners (who may be forced back into private rental housing i f they lose the ability to re-pay their mortgage debt). In light of the individual and societal burdens that homelessness implies, monitoring the risk of its incidence and attempting to understand its causes are a natural research target for academics and policy makers alike. But housing markets and individual housing careers are also characterised by tenure shifts in the opposite direction: homelessness is not always a long-term condition and households often re-integrate themselves into rental housing within a few weeks of losing their home (Ringheim, 1990; Urbina, 2006); similarly, homeownership as an institution is maintained by the thousands of tenant households that are able to move into owner occupancy every year. Yet the Core Housing Need and INALH measures alone do not seem to provide any indication of the possibilities (or the magnitude of the barriers) to upward tenure mobility. If one is to develop a more complex picture of housing consumption in all its dynamic modalities, then financial accessibility (or 'entry') is an area of knowledge that deserves as much attention as the more traditional conception of affordability currently favoured by Statistics Canada and C M H C . (To their credit, both agencies publish average housing prices and rents as well as changes in a derived housing affordability index. Unfortunately, the latter is only reported at the aggregate level, which prevents readers 29 from understanding the magnitude of the problem in terms of numbers of affected households.) As a concept distinct from affordability, 'inaccessibility' seeks to expand the debate through semantic appeals to the subjectivity of housing outcomes, instead of relying exclusively on the seemingly objective market assessment that the concept of affordability denotes. In other words, inaccessibility seeks to emphasise the household itself in relation to the market, as opposed to focusing on the market in relation to the household. This is more than a subtle difference in emphasis. While it is helpful to know that the average shelter cost to income ratio of tenants in Canada was 28 percent at the time of the 2001 Census (CMHC, 2005), researchers may also want to find out what is the proportion of households that do not meet financial thresholds for housing access (such as the 30 percent of income ratio required from mortgage applicants by banks). Users of the accessibility framework expect an appraisal not of the housing market per se, but rather of the individual households that participate in it. For this reason, academics and policy makers have discussed the importance of identifying not only income thresholds of affordability, but also entry costs or price levels at which participation in the private rental housing market or the transition from tenancy to homeownership become financially accessible (Burke and Hayward, 2001; G V R D , 2006; Hackworth and Wyly, 2003; Listokin et al, 2002; Y i Tong, 2004). A series of measures, genetically called 'entry' measures, can be employed to this end. (A good overview of this type of measure can be found in Gabriel et al, 2005). In the following chapters I will apply one version of these measures to the analysis of entry into homeownership, using 2001 data for households in Montreal, Toronto and Vancouver, 30 three markets which at the time were on the verge of entering the price boom of the early new millennium. The term 'accessibility' also denotes a series of non-monetary aspects, which the study at the centre of this thesis admittedly does not address. Indeed, the notion of accessibility encourages researchers and policy makers to take into consideration other factors (beyond income) that limit the ability of a household "to secure housing that is commensurate with its basic needs" (Gabriel 2005: 7). Such factors include a variety of circumstances that impinge on the way that households experience their housing conditions—what is commonly referred to as housing stress. Issues that impact the degree of housing stress include its temporal character (is it cyclical, episodic, or ongoing?), as well as the incidence of overcrowding, the quality and accessibility of facilities and amenities within and near the home, and security of tenure. M y concern here is specifically tied to the financial ability of Canadian metropolitan households to 'move up the tenure ladder.' The entry measures that I will be employing in my analysis do in fact make use of the 30 percent cost-to-income benchmark. However, they do so only in so far as it constitutes a widely used means test that mortgage lenders apply to would-be home buyers as a qualifying requirement for the allocation of mortgage financing. In theory, the entry measures used in this thesis could easily be calculated assuming the non-existence of qualifying cost-to-income standards. To reflect the differences between my approach and the official measures of affordability developed by C M H C (and Statistics Canada), I will use the terms 'financial accessibility' and 'capacity of entry' for the remainder of this thesis, except in those cases where I am referring to the broader set of issues that are generically termed 'affordability.' 31 C H A P T E R 2: A N A P P R O A C H T O E S T I M A T I N G I N T E R - M E T R O P O L I T A N CAPACITIES O F E N T R Y INTO H O M E O W N E R S H I P Vik Adhopia is a radio reporter with the Canada Broadcasting Corporation in Vancouver. In 2003, worrying that rising real estate prices would soon make it impossible to buy a first home, he and his wife set out to look for a house they could call their own. Having secured $50,000 for a down payment, the couple felt confident about the possibility of beating Vancouver's booming housing market and obtaining a mortgage on a small house with a backyard where their child could play. Having found an 800-square-foot dwelling that met their basic requirements, the couple placed an offer for $350,000. The highest bidder, however, offered $380,000. "It was such a frenzy at the time that people were putting offers without conditions," Adhopia explained. "At that point, we realized it was too late to buy a house in Vancouver." Almost three years later, Adophia and his family are still living in private rental housing in Vancouver. Their story was reported in an article in Vancouver's entertainment weekly The Georgia Straight in March of 2006 (Smith, 2006: 47). The article also mentions "a recent report" by the Greater Vancouver Regional District (GVRD)—the metropolitan area's administrative district authority—on the inaccessibility of home purchasing, and quotes from it as follows: "The income i thresholds required to own housing preclude home ownership for a large range of middle-income working people" such as "police officers, nurses, and teachers." The report is said to have found that "only six percent of renters could afford a single-family dwelling in the region, and just 11 percent had sufficient incomes to buy a townhouse," assuming a 10-percent down payment and a 4.4-percent interest rate. 32 These are astonishingly low percentages, even i f one considers that 63.5 percent of households in the Vancouver C M A were already homeowners in 2001. Everything else held constant, these figures imply that even i f every tenant household that could afford to buy a townhouse in 2001 had done so, almost one third of the region's total households would still have been unable to access homeownership. It is worth noting that in 1999, when the region's house prices relative to incomes were lower than during the current boom, a sharp income and wealth disparity was already setting apart the average Canadian homeowner from her tenant counterparts (Hulchanski, 2004). Charlie Smith's (2006) Georgia Straight article makes no mention of any demographic factors that the G V R D study may have taken into account. For example, one would expect that certain proportion of these households are headed by individuals at an early stage in their life cycle, earning 'starting-level' wages that are likely to improve after a few more years in the workforce. Presumably, a number of these young households will eventually reach sufficient income levels to afford homeownership. What share of households with insufficient income is headed by individuals in this age group? The magazine article leaves this and many other questions about that G V R D study unanswered. How do Vancouver households compare against other cities in Canada? What is the timeframe of the study? And how reliable are its data sources? Unfortunately, the report is not available to the public. This thesis seeks to reproduce (and perhaps expand on) this publicly unavailable G V R D study, based on what little is known about it from Smith's (2006) magazine article and from a summary document available on the GVRD's website (GVRD, 2006). Specifically, I will attempt to situate the case of tenants like Vik Adhopia in a 33 comparative socio-spatial context. Using the year 2001 as temporal basis, I will analyse the share of tenant households who had already been 'priced out' of the owner-occupied-dwelling markets in Canada's three largest metropolitan areas. A demographic and socio-economic profile of these households will be constructed, to help us understand the causes and effects of this inaccessibility problem. The present chapter explains my approach to deriving those estimates and to producing those household profiles. One reason for choosing 2001 as the reference year for my study is that prices rose for the most part at a much higher rate during the three-year period of 2002-2004 than they did between 1996 and 2001, particularly in Montreal and Vancouver (Figure 7). My analysis therefore evaluates the capacity of entry into homeownership of Figure 7: Annual change in average house prices, selected metropolitan areas, 1996-2004 (Percentage change) • M ontreal EI Toronto • V a n c o u v e r 1996 1997 1998 1999 2000 2001 2002 2003 2004 Note: Based on average annual sales prices as per CREA's M L S listings for each metropolitan area, reported by C M H C and adjusted to 2004 constant dollars by the author. C R E A ' s definition of Metropolitan Area differs from Statistics Canada's definition of CMAs . Sources: CMHC/Canadian Real Estate Association; Statistics Canada. 34 metropolitan households as it stood in the year before dwelling prices embarked on their explosive assent. The resulting snapshot not only provides some idea of which are the households most likely to be affected by the price boom, but also creates a basis for comparative cross-sectional analysis with other reference years (an undertaking that is beyond the scope of this thesis). A second reason for choosing 2001 as my reference year is the availability of pertinent data. The income and demographic records used in this study come from the 2001 Census Public Use Micro-data File (PUMF) on Individuals. In Canada, a census of population is conducted every five years, and the most recent published results are for the one conducted in 2001. To my knowledge, no other source of data as detailed and with a sample size as large as the Census PUMF is readily available for years between each Census. The nearest possible source is Statistics Canada's annual Survey of Household Spending, a dataset that has unfortunately a much smaller sample size (only 21,000 households nationwide) and does not provide as much socio-demographic detail about the respondents. The lack of alternative data sources on income distribution and population growth (disaggregated by a sufficiently large variety of demographic characteristics) for the period 2002 to 2004 explains why I could not focus directly on the changing inaccessibility picture during the price boom years. Had a household been in the same financial situation in 2001 as Adhopia's was in 2003, would it have been economically able to buy a moderately priced house then? This study cannot answer the question with certainty, i f only because home purchases depend on more than household balance sheets and income statements. However, my results and 35 analysis should provide an indication of the degree of financial inaccessibility that pre-existed the onset of the price boom. Does uneven access to homeownership matter? In a study of income and wealth differences by tenure, Hulchanski (2004) suggests that access to homeownership is one of the main wedges prying open the income and wealth gaps in Canada. Differential access to homeownership directly or indirectly produces and reinforces inequality through a host of inter-related socio-economic factors, which include demographic trends, rates of migration, income levels, demand-related policies, economic growth, interest rates and credit availability, and so on (Bourne, 1981: 75; Hulchanski 2005). Four key ways in which some of these factors interact with tenure-related inequality are worth mentioning here. First, while not all homeowners are wealthy, enjoy high incomes, or benefit equally from ownership (Carter, 2004; Forrest et al, 1990; Retsinas and Belsky, 2002), aggregate statistics show that the amount of equity on a dwelling is the main source of household wealth in Canada (Hulchanski, 2004: 84). Tenants generally lack the same opportunities for net asset accumulation and financial security that many homeowners derive from the appreciation of their property over time. This type of inequality can have consequences that extend beyond the present, because wealth disparities produced by this opportunity differential may be transmitted from one generation to the next (Boehm and Schlottmann, 2002, 1999). Indeed, whereas homeowners who have not extracted all of their housing equity before death may choose to transfer their wealth to their children, tenants generally lack dwelling-related assets that their children could potentially inherit. 36 Secondly, tenants may experience a substantial reduction in their standard of living upon retirement. The Canadian state pension system is designed mainly as a complement to individual savings, and regular payments are equivalent to only a fraction of the recipient's pre-retirement employment income. Unable to rely on wealth accumulated through homeownership, tenants who lack savings or non-financial assets to capitalise may be forced to significantly adjust their budgets after retiring from the work force. Thirdly, low-income tenants are considered to be at a greater risk of experiencing some form of homelessness (Ringheim, 1990). This is particularly the case for tenants who have little or no savings, because an income interruption (caused for example by a layoff or illness in the family) may prevent them from being able to pay their rent (Hiebert et al, 2005). And as The New York Times reports in a recent article (Urbina, 2006), unemployed individuals experiencing difficulty in maintaining or accessing housing can have a more challenging time with finding a new job. Finally, the case of immigrants deserves special attention, i f only because newcomers now constitute approximately two-thirds of Canada's population growth (Statistics Canada, 2006). Immigrants settle predominantly in the nation's three largest metropolitan centres, and they are therefore an important and rapidly growing segment of housing demand in these cities. But the journey that immigrants typically undertake entails a period of personal adjustment that often places them—at least in the first years of settlement—in a comparatively disadvantageous situation with the native-born, especially with regards to the labour and housing markets. We know for example that recent immigrants are more likely to be tenants and experience core housing need than 37 the native-born. However, we also know that homeownership rates for immigrants with more than 15 years in Canada.are higher than the non-immigrant population's (CMHC, 2004), suggesting not only economic improvement over time, but also a strong preference for owner-occupied housing. Based on recent and historical statistics for immigrant tenure patterns over time, scholars and policy makers have come to regard homeownership as a reliable indicator of the degree of social and economic integration of the non-native-born. Immigrant homeownership is widely seen as representing economic success and a serious commitment to settling and investing in the new home country (Alba and Logan, 1992; Haan, 2005a; Ray and Moore, 1991;,Skaburskis, 1996). Homeownership, however, is more than simply an expression of consumer preferences in the fulfilment of individual desires. The propensity and aspiration to own one's house is also a function of cultural, market, and policy forces that change across time and place (Kemeny, 1981). By privileging an uncritical assessment of past observations at the expense of the locally specific material and ideological contexts in which tenure preferences develop, this deeply seated belief inadvertently makes the notion of citizenship (or the commitment to a given community of belonging) symbolically dependent on proprietorship. One of the dangers here is that in times of economic instability, falling rates of immigrant homeownership—resulting from a decline in the financial accessibility of housing-might be interpreted in some quarters as a lack of commitment to Canada, which could contribute to the development of negative public attitudes towards immigration. Furthermore, in as much as homeownership "serves as a substantial base upon which immigrants can further flourish economically and culturally" (Myers and Yang Liu, 38 2005: 351), its inaccessibility can hamper their personal development and settlement process, and thus have far-reaching consequences both for newcomers and the receiving society. The social and individual risks associated with the financial inaccessibility of owner occupied housing are not trivial affairs. It is thus crucial to monitor the size and characteristics of the tenant population that is economically unable to realise an aspiration to homeownership. Knowledge of the demographic, geographic and socio-economic features of this segment of the population should help academics and policy makers understand the intricate ways in which tenure-related inequality can be magnified through interaction with other social and economic factors. Capacity of entry and household characteristics Recent Census-based studies of homeownership propensity in Canada—primarily designed to compare immigrant and non-immigrant households—have attempted to capture the effects of various demographic and socio-economic axes of household difference on tenure outcomes (Edmonston, 2004; Haan, 2005a; Lapointe Consulting Inc. and Murdie, 1996; Laryea, 1999). The independent variables used in such studies (summarised in Table 2) provide a template for the types of characteristics that this thesis will use in order to develop a profile of households already priced out of homeownership in 2001, the year before the onset of a house-price boom. Three of those variables are not included in my study. The first of these is employment status (the assumption behind this omission is that income is positively correlated with stability of employment, and that income levels—particularly above-average incomes—tend not to fluctuate significantly downwards until retirement; moreover, banks generally require a predictable flow of 39 Table 2: Variables used in recent North American studies on homeownership propensities VARIABLE: AUTHOR AND DATE: • Lapointe Consulting Inc. andMurdie (1996) Laryea (1999) Edmonston (2004) Haan (2005a) Sex* X Age* X X X X Household composition X X X X Presence of children X X Marital status* X X Household size x X Schooling* X X X Employment status* X Canadian-born* x X Ethnic category* X Multiple [ethnic] origins* X [National] citizenship* X Years since migration* x X X X Speaks English/French* X X Country/region of origin* x X Dwelling type x Household income x X X X Census year X X City of residence x X X X * Describes a characteristic of the "Primary Household Maintainer," which is the person that Census respondents identify as the one who normally contributes the most towards payment of the household's shelter expenses. future income, which typically is associated with stable employment). The second variable I left out of my study is dwelling type, simply because it is not available in microdata form as part of the 2001 Census PUMF file. The third and final variable I have left out of this research is country of origin, due to the poor level of disaggregation that the PUMF dataset provides. While these three information gaps are admittedly a drawback for my study, I have attempted to compensate by adding two variables not 40 taken into account in the four studies cited above. These are: visible minority status, and generational ties to immigration of the primary household maintainer. In this way, I hope to contribute new perspectives to recent homeownership research in Canada. From the technical point of view of the scholar, the unit of housing consumption is the household, not the individual per se. But in studies that engage on descriptive statistical analysis of households, researchers have to contend with the fact that the demographic characteristics of individual household members are not uniform, while the household, as an object of statistical description, needs to be so. Other than developing multivariate models and methods of statistical analysis that can take into account a given set of characteristics of each member of a household, researchers have little choice but to report on the characteristics of the main contributor to the household's shelter expenses, typified in the Census of Canada as "Person 1" or the "Primary Household Maintainer" (PHM). Marion Steele (1979) proposes an economist's rationale for such an approach: i f housing is the result of a consumer's choice, then there must be a hierarchy of decisions leading to it. "The first of these is whether or not to occupy a separate dwelling unit. This is equivalent to the decision whether or not to head a separate household" (cited in Doucet and Weaver, 1991: 306). Aside from the well-known problems associated with the 'rational choice' school of thought, this argument is undermined by the growing presence of double-income households in Canada's metropolitan areas, because more than one person may be contributing significantly to cover the contemporary household's shelter expenses (Buzar et al, 2005; Ley, 1999). Admittedly, my thesis does not attempt to solve the methodological challenge associated with the selection of a unit of analysis. 41 Instead, I have opted for the standard approach of reporting on the characteristics of the person who was identified by census respondents as the primary contributor to housing expenses, acknowledging the bias that it introduces. The 2001 Census PUMF (Individuals File) dataset that was accessed for researching this thesis was the updated version of April 26, 2006. It contains non-aggregated data based on a 2.7 percent sample of anonymised responses to the long form Census questionnaire. (While Statistics Canada attempts to reach all households to conduct the Census, the long form questionnaire was sent to only one in five enumerated households in 2001. As its name implies, the long form contains an additional set of questions that are not part of the regular short form.) The sample size or total number of unweighted cases in the PUMF (Individuals File) is 801,055 for all of Canada. One hundred and forty variables are available for cross-referenced analysis. What follows is a brief description of the variables that I have used in the present study: - Age (Person 1): The age of the Primary Household Maintainer (PHM) provides some indication of the life-cycle stage of each household at the time of the 2001 Census. Correspondingly, only PHMs above the age of 15 were considered in this thesis. The age group breakdown was coded as follows: 15-24; 25-34; 35-44; 45-54; 55-64; 65-79; 80 and over. Household structure: Five categories of this variable were employed in this thesis: Lone parent families; Couples; Multiple family households; Non-family households of two persons or more; and One-person-only households. - Presence of children in the household: The household structure variable discussed above was broken down in order to take the presence of children in the household 42 explicitly into account. Five categories were set up: Lone parent families; Couples with child(ren); Multi-family households with child(ren); Couples without child(ren); and All other households without child(ren). Marital status (Person 1): The household structure variable (discussed above) has been recoded again in such a way as to allow an assessment of differences in capacity of entry into homeownership according to marital status. Five categories were used: Married, no children; Married, with children; Common-law, no children; Common-law, with children; and All other households. Highest level of schooling (Person 1): Five subcategories are used for this variable: Less than high school grad; High school grad; Trades certificate/diploma; Some post-secondary; and Bachelor and above. Language (Person 1): For this household characteristic, I combined the two PUMF variables "Language spoken at home" and "Knowledge of official language(s)." I am therefore able to provide information on households whose P H M : Speaks an official language at home; Speaks a non-official language at home, and knows English and/or French; and Speaks a non-official language at home, and does not know English and/or French. Citizenship (Person 1): The PUMF file provides three categories for this variable. c Individuals who are citizens both of Canada and another country are counted twice, the second time being in the category "Other citizenship." For the present study, I have created five sub-categories for this variable: Canadian citizen by birth; Naturalised Canadian citizen; Other citizenship/less than 10 years in Canada; Other citizenship/10 years in Canada or more; and Non-permanent 43 resident. Households 'headed' by a person who is a Canadian citizen by birth but also holds citizenship in another country are counted only once in this variable, under 'Citizen by birth.' The same is true of dual citizens who are Canadian by naturalisation. Sex (of Person 1 and of Spouse): In this thesis, I have used this variable to describe a household based on the sex of both Person 1 and of his or her spouse or common-law partner (where applicable). Consequently, the totals for this variable do not add up to the total household counts for each city. In the case of couples, subtotals for the categories 'male' and 'female' (these are the only responses that can be given in the Census questionnaire) tally both members of the relationship. Visible minority (Person 1): The Employment Equity Act in Canada defines visible minorities as "persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour" (Statistics Canada, 2003: 143). While the 2001 Census questionnaire does not mention the term visible minority, it asks the question "Is this person:" and then provides 11 written categories plus the option "Other - Specify." The categories are labelled for the most part with an adjective denoting regional, national or linguistic characteristics such as Latin American or Japanese, but include also White.and Black. Respondents were instructed to mark or specify all the categories that apply, providing the possibility to self-identify as belonging to multiple groups. In the 2001 Census PUMF dataset, this variable is coded with only five categories: Chinese; South Asian; Black; Other visible minority; and Not a visible minority [i.e. White]. The 44 first three categories correspond to the largest visible minority groupings in Canada. I have retained the five categories in my analysis. Ethnicity (Person 1): While visible minority is an individual characteristic, the ethnicity variable describes the "ethnic or cultural group(s)" (not citizenship or nationality) to which the individual's ancestors belong. The 2001 Census questionnaire provides a small list of examples and leaves four blank spaces to specify alternative responses for each household member. For the purposes of this study, I have collapsed the numerous categories provided in the PUMF dataset into 11 groups. I have cross-tabulated these categories with the variable "Immigration Status" in order to generate a profile of ethnic backgrounds for immigrants only. I use nine ethnic groupings: European; African; Arab; West Asian; South Asian; East and Southeast Asian; Latin, Central, South American; Caribbean; and Multiple ethnicities. Immigration cohort (Person 1): Six immigrant cohorts, based on year of arrival, are used in this study: Arrived before 1961; 1961-1970; 1971-1980; 1981-1990; 1991-1995; and 1996-2001. Generation (Person 1): The term 'first generation' is used by Statistics Canada as a demographic category to designate immigrants who were born outside Canada. 'Second generation' applies to individuals who were born in Canada of at least one parent who was an immigrant born outside Canada. Individuals born in Canada with at least one grandparent or a more distant ancestor born outside Canada are categorised as 'third generation and over.' In this thesis, I use the four categories that are provided in the PUMF dataset for this variable: First 45 generation; Second generation/both parents born outside Canada; Second generation/one parent born outside Canada; and Third generation and over. - Household income: In the PUMF dataset, this variable reports before-tax household income for the calendar year 2000, broken down into 20 or so income ranges {e.g. $0-$4,999; $5,000-$ 10,000; and so on). A detailed breakdown of the recoded income ranges used in this study for each C M A will be provided below. The household income variable forms the basis of the capacity of entry analysis in this thesis. Using this variable, I estimate the income level that a household would require to qualify for a mortgage on an average-priced or, alternatively, a moderately priced house in each metropolitan area (my definition of 'moderately priced' is provided later in this chapter). For the purposes of this study, tenant households that earned incomes below their C M A ' s qualifying threshold in 2000 are considered to have been unable to access homeownership in 2001. For example, this study will state that a Vancouver tenant household reporting a total annual income of less than $85,000 in the 2001 Census would have found itself below the mortgage-qualifying threshold for an average priced dwelling in that year. The capacity of entry estimates were produced using less than perfect information, thus requiring the adoption of a number of conservative assumptions. Some of these are borrowed from Y i Tong's (2004) recent study of homeownership affordability in metropolitan U.S.A. In that study, the author employs a homeownership affordability index to compare in time and space the gap between median family income and the income required to qualify for a mortgage on a median priced home. Y i Tong's index is expressed as a percentage of income: 46 For instance, a ratio of 130 indicates that the median-income family needs to have 30 percent more than its actual income to qualify for a mortgage for a median-priced home. In contrast, a ratio of 90 indicates that the median-income family needs to use only 90 percent of its income to meet the lender's income requirements for purchasing a median-priced home. (YiTong, 2004:14) In this thesis, I will alter Y i Tong's homeownership affordability index to produce an alternative measure, expressed instead as a percentage of a given population. The main difference between the measure that I have adopted and the one used by Y i Tong stems from the type of information that it makes readily available to the reader. For example, a ratio of 130 under Y i Tong's formulation simply tells the reader that households with incomes that were 30 percent below the median would not have qualified. It would take someone with access to income and population data to be able to calculate the corresponding number of households that fit such income characteristics, and thus arrive at the same estimates of household numbers as the ones I have produced in my study (except of course that my data are expressed as a function of average rather than median figures). The key advantage of the version of the index that I am adopting here is that it does not require readers to access any data or to conduct additional calculations, because the percentage that serves as a measure relates to the proportion of tenant households who are negatively affected by high dwelling prices. Being expressed as a percentage, the statistic is equally helpful as a ratio or other type of index when it comes to conducting comparisons between different cities and different points in time. The measure that I am employing here does require the researcher to obtain more detailed income data than what Y i Tong employs, but it does not impose any data access requirements on users who are simply interested in obtaining frequency counts or proportions of estimates of affected households. 47 For the purposes of this thesis, I assume that tenant households were not owners of dwellings they did not occupy, and that they would have desired to purchase an average or moderately priced house in the C M A where they lived in 2001. Adapting Y i Tong's (2004) study to the Canadian context, I also assume that all tenant households could put together enough money to make a 10 percent down payment, that all households would have otherwise been able to meet all the credit requirements of a lending institution, and that they would have received an interest rate equivalent to the average chartered-bank-administered five-year conventional mortgage rate as published for May 2001 by the Bank of Canada. (I use the May 2001 rate because the Census was conducted in that same month. Furthermore, I applied the conventional mortgage rate in recognition of anecdotal reports that banks are offering similar rates to households with either 10 percent or 25 percent down payments.) Annualised monthly payments were calculated using an Internet-based mortgage payment calculator accessed through the website of a Canadian chartered bank. I also assume that the qualifying yearly household income required by the lending institution is based on annualised monthly payments equivalent to 30 percent of the household's annual income. In this thesis, payments include both interest and principal in the case of the mortgage (calculated on the basis of a 25-year amortisation period), as well as an amount equivalent to the 2001 annual average of household expenditures on mortgage insurance, property tax, homeowners insurance, maintenance and repairs, and water and heating costs for C M A s of one million people or more, as estimated by Statistics Canada from the 2001 Survey of Household Spending (SHS, Table 62F0035). It is unfortunate that Statistics Canada's publicly accessible tables on household shelter expenditures are only compiled in this aggregate 48 format instead of in detailed breakdowns for major CMAs. While the average non-mortgage shelter costs I used here are likely to vary by about two hundred dollars a year between Montreal, Toronto and Vancouver (see Statistics Canada, 2003, Table 7), this small difference would only add or subtract an immaterial $600 to $700 a year to the qualifying annual household income. For this reason, I have retained in my calculations the publicly available expenditure figures as provided by the SHS. The house price data used in this study also require some discussion. While the Census collects self-reported house price information and makes it available in the PUMF file, I have not used that particular Census variable in my analysis. The 2001 Census long form questionnaire asks homeowners to state how much they would expect to receive for their dwelling i f they were to sell it at the time when the Census was conducted. It is unlikely that the amounts stated by respondents were consistently based on recent professional appraisals. The data are therefore not necessarily reflective of the actual market resale value of dwellings at that time. Moreover, the maximum dwelling value tabulated in the 2001 Census PUMF (Individuals file) dataset is $200,000; anything higher is simply recoded by Statistics Canada and labelled "$200,000 and over." Given the much higher prices at which dwellings were selling in Vancouver and Toronto in 2001, the average price figures obtained through this variable are incongruously low, making PUMF data inadequate for housing price analyses in these expensive markets. A second option for house price data is published by C M H C (2005). These aggregate data are expressed as yearly average prices per metropolitan area, and are based on the Canadian Real Estate Association's M L S listings. MLS stands for Multiple Listing Service, the trademark name of the Canadian Real Estate Association's database 49 of new and resale residential properties that are listed with participating realtors. The price information reflects the reported amounts for which listed properties sold in the market. The data are therefore not based on a census of all dwellings or even on a scientific sampling of all house sales in a given city. Average price is calculated by dividing the total annual dollar amount of sales by the total number of annual sales. Moreover, Statistics Canada's definition of Census Metropolitan Areas does not correspond exactly with the metropolitan boundaries used by the Canadian Real Estate Association and local real estate boards. Despite these limitations, I have opted for CREA's MLS statistics because of the large number of observations from which they are produced. My choice is also supported by the widespread use of these aggregate data in the public and private sectors (examples include C M H C , 2005; TD Economics, 2005). The MLS price data were used in conjunction with PUMF income and tenure data in order to create three categories of tenant households: those who would have qualified for a mortgage in 2001 under the conditions laid out above (Group 1), plus two categories of those would have not (Group 2 and Group 3). Group 2 corresponds to tenant households with incomes that fall below the threshold required to qualify for a mortgage on an average priced house, but who would qualify for a 'moderately priced' unit (meaning a dwelling priced at just a fraction of the average price; a precise definition of what I mean by 'moderately priced' in this study is provided below). In turn, Group 3 represents tenant households who would have been unable to qualify for a mortgage on either of these two differently priced dwellings. The continued tenancy status of households in Group 1 can be attributed to a number of reasons that I will not attempt to explain in depth in this thesis, but which 50 include: a lack of savings to afford a down payment (as in the case of recent graduates entering well paying jobs) or a 'lifestyle' choice to not enter homeownership (Varady and Lipman, 1994), but also a real or perceived lack of job security due to the growing precariousness of today's labour markets, an expectation of future departure from place that discourages investment in real property (see for example Owusu, 1998), and the unavailability of units deemed appropriate to the household's needs (as in the case for example of large households). In addition, as I will discuss later in this paper, groups that identify with certain visible minority or ethnic background categorisations display more of what appears to be a strong preference for homeownership compared to other similarly categorised groups, which could mean that the opposite is true for some of these households. Finally, in the case of immigrant households, especially recently arrived ones, a possibility is that much of the combined household income is sent abroad as remittances (DeVoretz, 2004), thus depriving the household from the ability to save enough income to afford a down payment. In turn, Group 2 is made up of households that may be able to get a foot in on the low end of the housing ladder but are perhaps waiting to improve their financial situation in order to afford a better-located and better-appointed housing unit. Alternatively, households in Group 2 may find themselves in some of the same situations I have listed for Group 1. Finally, households in Group 3 represent tenants with the lowest household incomes relative to house prices in their respective C M A , such as students, elderly life-cycle renters (Varady and Lipman, 1994) who depend on fixed incomes, individuals and families on social assistance, and the working poor. Households in this third group are the main focus of my research. 51 My decision to tabulate the homeownership entry capacity of the first and second groups of tenant households is not inconsistent with my interest in developing a profile of households belonging to Group 3. The decision arose from recognition of the wide range of prices that tend to be masked by average figures within and between cities: a 750 square-foot low-rise condominium in the Vancouver suburb of Port Coquitlam will command only a fraction of the price of a waterfront two-storey house in the city's Kitsilano neighbourhood or a heritage mansion in Toronto's upscale Rosedale area. Moreover, average prices do not always match the actual amount for which the dwelling is sold, as one of the largest real estate companies in Canada recently reported (ReMax, 2005). In fact, my own analysis of MLS data confirms this realtor's findings. Table 3 shows the breakdown of MLS sale prices for all single dwellings (including detached, attached, and condominium-type units) in 2001 for metropolitan Montreal and Toronto. In Greater Montreal, only 39.9 percent of MLS-listed sales were closed at a price of at least $125,000 (the average metropolitan price was $128,000). Similarly, only 35.5 percent of listed sales in metropolitan Toronto were recorded at prices of at least $250,000 (the local average price being $251,500). As these two examples show, average figures are not necessarily good reflections of moderate house prices, and may in fact overestimate the numbers of non-qualifying tenant households. In an attempt to "use the behaviour of recent homebuyers as a critical benchmark'' (Listokin et al, 2002: 21), I approximate what might be regarded as a 'moderate price' for a dwelling in the 2001 metropolitan markets of Montreal and Toronto, by analysing cumulative sales-volume-per-price figures and identifying the minimum price recorded by approximately two thirds of all MLS listed sales (Table 3). For both of these C M A s , 52 Table 3: Determining metropolitan 'moderate price' figures for 2001 Montreal Toronto Vancouver Listed selling price: Number of units % Listed selling price: Number of units % _ ... . Range of June 2001 Dwelling type ,. a . . . 3 l r benchmark prices At least $250,001 1,518 6.7 At least $400,001 5,387 8.4 2-BR apartment $83,590 To $297,352 At least $200,001 2,632 11.7 At least $300,001 13,571 . 21.0 3-BR attached $141,745 To $287,231 At least $150,001 5,679 25.2 At least $250,001 22,875 35.5 3-BR detached $215,448 To $705,907 At least $125,001 8,985 39.9 At least $190,001 41,538 64.4 At least $100,001 14,518 64.5 At least $160,001 51,468 79.8 (Based on MLS Housing Price Index At least $80,001 19,579 87.0 At least $130,001 59,228 91.8 data for May 2006) At least $50,001 22,187 98.6 At least $90,000 59,861 99.0 Total units sold 22,501 100 Total units sold 64,509 100 Average 2001 price: $128,500 Minimum selling price of approximately two thirds of all MLS-listed units sold in Greater Montreal in 2001: $100,000 (includes detached, attached and condomimum-type dwellings) Average 2001 price: $251,500 Minimum selling price of approximately two thirds of all MLS-listed units sold in Toronto in 2001: $190,000 (includes detached, attached and condomimum-type dwellings) Average 2001 price: $285,900 Low-end estimate of benchmark price for three-bedroom detached units in Greater Vancouver, June 2001: $215,448 Adapted from: CMHC/Greater Montreal Real Estate Board; CMHC/Toronto Real Estate Board; Real Estate Board of Greater Vancouver. this is equivalent to approximately 75 percent of the average price: $100,000 in Montreal and $190,000 in Toronto. In the case of Vancouver, similar data are not available; instead, I have simply applied the 75 percent weight to its metropolitan average, which yields a 'moderate price' figure of $215,000. Curiously, the Real Estate Board of Greater Vancouver's (http://www.realtylink.org/statistics/buyers_ market_HPIjnain.cfm) M L S Housing Price Index estimate for the June 2001 benchmark price of a three-bedroom detached dwelling at the low-end was also equivalent to approximately three quarters of the average 2001 M L S price for the city-region (Table 3). It should be noted, however, that the R E B G V benchmark does not reflect prices of existing dwellings, but instead constitutes a composite of realtors' estimates of the average amount of dollars that consumers spent on various services and facilities of a dwelling, such as master bedrooms, bathrooms, recreation facilities, and so on. While the low-end benchmark 53 prices for attached dwellings and apartments were considerably lower than the $215,000 I have chosen as the C M A ' s moderate price figure, this does not imply that significant numbers of units were available for sale at or near such prices during the reference period of June 2001. In any case, the mere existence of lower priced units does not necessarily mean that those prices are representative of what was primarily available on the market. The sales volume-by-price distribution in Montreal and Toronto suggests that the $215,000 price is a more reasonable estimate of what households could expect to pay for a moderately price dwelling in metropolitan Vancouver in 2001, holding constant the C M A ' s equilibrium price level for average and moderate units. Table 4 summarises the average and moderate mortgage-qualifying household income thresholds in each C M A , including the figures I used in their calculation. In general, qualifying incomes did not fall neatly at the low or high end of the ranges provided in the PUMF dataset. I have therefore used the lower limit of the PUMF income range that captures the qualifying income in each C M A to define the minimum income of Group 1 households, and applied the similar adjustment-to define Group 2 and 3 (Table 5). For example, the qualifying income for the average-priced house in Toronto was $79,300, which falls within the PUMF income range of $75,000 to $84,999. A household was considered to be at or above the average-mortgage qualifying threshold i f its annual income was $85,000 and over. The only exception relates to Vancouver's moderately priced dwellings, for which the qualifying income of $69,350 was less than a thousand dollars below the high end of the corresponding PUMF income range. In this case, the qualifying threshold has been set at $70,000. 54 Table 4: Mortgage-qualifying income thresholds, before tax, 2000 (In nominal dollars; 2001 average house prices and shelter expenses) Montreal CMA Toronto CMA Vancouver CMA Average house price $128,550 $251,500 $285,900 Moderate house price $100,000 $190,000 $215,000 Estimated yearly payments: Mortgage, Group 1 (average-priced house) $10,375 $20,300 $23,075 Mortgage, Group 2 (moderately priced house) $7,800 $15,250 $17,300 Property taxes $1,530 $1,530 $1,530 Homeowners' insurance $300 $300 $300 Water, fuel and electricity $1,595 $1,595 $1,595 Mortgage insurance $75 $75 $75 Qualifying income threshold1, households in Group 1: $46,350 $79,350 $88,350 Qualifying income threshold1, households in Group 2: $37,650 $62,500 $69,350 Average mortgage rate, May 2001: 7.75% Based on annualised payments using the average chartered-bank-administered five-year conventional mortgage rate for May 2001, with an amortisation period of 25 years, a 10 percent down payment, and monthly payments of interest and principal. Payments also include yearly averages for mortgage insurance, property tax, homeowners insurance, and water and heating costs are for 2001, for C M A s with populations of one million or more (for an explanation of why the latter set of figures is the same in all three cities, see page 49 in this thesis). Figures have been rounded. Sources: C M H C (2005); 2001 Survey of Household Spending (Table 62F0035); Bank of Canada (CANSIM Table V122521). Table 5: Before-tax household income ranges, 2000, Groups 1, 2 and 3 (In dollars, not adjusted for inflation) Tenant households: Before-tax household income, 2000 Montreal CMA: Toronto CMA: .Vancouver CMA: Group 1 (qualifies for average mortgage) $45,000 or more $75,000 or more $85,000 or more Group 2 (qualifies for moderate mortgage) $35,000 to 44,999 $60,000 to 74,999 $70,000 to 84,999 Group 3 (does not qualify for either mortgage) $34,999 or less $59,999 or less $69,999 or less Group 1 and 2 represent tenant households with sufficient pre-tax household incomes to qualify for a mortgage on an average-priced dwelling and a moderately priced dwelling, respectively. Group 3 encompasses all tenant households who would be financially unable to qualify for a mortgage on either of those two differently priced types of dwelling. Figures adapted from the household income variable of the 2001 Census P U M F - Individuals file. 55 An analysis of these multiple qualifying thresholds, expressed in terms of number of tenant households that fall below or above them, will be conducted in Chapter 3. For now, I will simply comment that as in the case of dwelling prices, qualifying thresholds exhibit an uneven geography across the three metropolitan areas, with Vancouver posting the highest income requirements, followed by Toronto and finally Montreal, which sits considerably behind the other two cities (Table 5). But before moving to that chapter, I will review some general background information pertaining to each of these three CMAs, which should provide the reader some indication of why differences in prices, household incomes, and qualifying thresholds have much to do with the particularities of each metropolitan area's history of development, population distributions, mix of units in the housing stock, and so on. Capacity of entry and metropolitan geography Understanding the diversity that characterises cities and gives them their particularity is not an easy task. Descriptive statistics continue to play an important role in these endeavours, but it is helpful not to rely exclusively on the highest levels of aggregation. Studies of homeownership as a metropolitan institution are a case in point. Figure 8 provides a geographical illustration of the proportions of Canadian households that are tenants and homeowners in each of the six largest C M A in the country in 2001. At first glance, the 2001 distribution of owner occupancy shows a remarkable degree of heterogeneity across most Canada's largest metropolitan centres. Four of the six cities had household homeownership rates at par with the Canadian overall rate of 60 percent. Only Calgary (with a rate of 70 percent of households) and Montreal (with a rate of 49.9 56 Figure 11: Household distribution by tenure, Canada's six major metropolitan centres, 2001 Note: Total number of households in brackets. Source: Census 2001, PUMF (Individuals file); Shape files: Statistics Canada, Cartographic Boundary Files, 2001 Census 2nd Edition, 92F0166XCE Census Metropolitan Areas/Census Agglomerations, January 2001. 57 percent) were noticeably different from the rest. But as I will show in the next two chapters, the tenure landscape of metropolitan Canada is in fact more spatially contingent than this map suggests; its high level of aggregate representation masks a considerable amount of variability, such as the one that characterises various household types and their financial ability to move from residential tenancy to homeownership. To help understand this variability, I will now provide a brief overview of the three CMAs that constitute the sites of this study. Montreal CMA Reviewing the past few years of quarterly press releases from the Chambre Immobiliere du Grand Montreal, one can get a sense of the extent to which rampant gentrification, a boom in new residential construction, and an upsurge of redevelopment and renovation projects have transformed the city-region over the past ten years, following a prolonged period of housing market stagnation that lasted well into the late * 1990s. High unemployment and widespread concern over the potential outcome of a 1995 referendum on the province's independence from Canada appears to have caused a glut of for-sale properties in the market, which a major realty company has blamed for the relative lack of interest in homeownership (ReMax, 2005). In 2001, when average house prices started to recover (by 2002 they were growing at the fastest pace in Canada), just over 50 percent of dwellings were owner occupied—a rate of homeownership that was approximately 10 percentage points lower than Toronto's and Vancouver's. But this rate differential was not only related to inventory surpluses and local incomes. As I will show later, Montreal had in 2001 the highest proportion of tenants with sufficient incomes to purchase an average or moderately priced dwelling. According to Brian Ray 58 Figure 9: Structural dwelling types, selected CMAs, 2001 (Number of dwellings) Adapted from: 2001 Census of Canada, Catalogue number 95F0327XCB2001004 (1999: 66-67), one must take into account the city's "distinctive culture of property based on a large number of small-scale landlords, and a large and varied stock of low-rise rental units ('plexes' of various configurations and low-rise garden-style apartments)." In 2001, the housing stock in Montreal continued to be dominated by apartments in buildings with four or fewer floors (Figure 9), and the proportion of single-family detached units as a percentage of the total housing stock was smaller than in Toronto or Vancouver. In terms of population, Montreal is the second largest C M A in Canada, although it has the highest population density per square kilometre (846.6 individuals). Between 1986 and 2004, the average annual rate of population growth in Montreal was below one 59 percent, which is also below the Canadian average and less than half the rate of Toronto and Vancouver (Heisz, 2005). Montreal receives fewer immigrants, who make up a smaller proportion of the total C M A population. Nonetheless, it has the third largest number of immigrant households in Canada, and the urban landscape in most parts of the city is characterised by its ethnic and cultural diversity (Germain, 1999; Ray, 1999). But the metropolitan geographies of immigrant settlement vary by immigrant class. Canada's immigrant selection process assigns different requirements for applicants to each entry category or programme. These include family reunification, skilled workers, business class, and refugees. Montreal receives proportionally more refugees (who are often economically deprived upon arrival) than Vancouver and Toronto, and considerably fewer newcomers in the business class (Ley and Smith, 2000; Chui, 2003). Montreal, Toronto and Vancouver together comprised just over half (53.5%) of the total labour force of all C M A s in 2001, and Montreal had the second largest share at 17.3 percent (Table 6). Montreal had the lowest participation rate among the three C M A s analysed in this study (65.7%), while Toronto had the highest (69.8%). Labour markets in the three C M A s are undergoing an important structural transformation (Heisz, 2005): in all three, the share of C M A employment in manufacturing fell between 1986 and 2001 while the business services sector made noticeable gains, a shift that reflects typical efforts of metropolitan centres seeking to achieve global city status (see for example Sassen, 2001). Toronto CMA Although it has only the second largest population density per square kilometre (793.3 individuals), Toronto is the country's most populous metropolitan centre. As of 60 2001, it had the largest proportion of households in Canada at 14.3 percent (Table 1), and the largest share of the labour force in all CMAs (Table 6). While Toronto's rate of unemployment in July 2001 (seasonally adjusted, three-month moving average) was only slightly below the aggregate rate for all CMAs, the average family income was higher than Vancouver's and Montreal's (Figure 3). It is then perhaps not a surprise that Toronto also has the highest rate of homeownership, as well as the largest stock of single detached dwellings, both in absolute numbers and as a share of its total supply (Figure 9). Toronto's residential landscape is also distinguishable by its large number of high-rise buildings, including those that were built in the 1960s and 1970s to provide public housing (Murdie, 1994; Ray, 1999). Table 6: Labour force characteristics by metropolitan area, July 2001 (Seasonally adjusted, three-month moving average) In thousands of individuals Percentage Proportion of labour force Partici- relative to Population Labour force Employ- ' ment Unemploy-ment pation rate Employ-ment rate Unemploy-ment rate that of all CMAs Toronto 3,948.3 2,756.4 2,584.5 171.9 69.8 65.5 6.2 25.7 Montreal 2,825.1 1,855.7 1,709.9 145.8 65.7 60.5 7.9 17.3 Vancouver 1,712.1 1,134.4 1,069.5 64.9 66.3 62.5 5.7 10.6 Ottawa-Gatineau 874.6 611.4 572.8 38.5 69.9 65.5 6.3 - 5.7 Calgary 788.1 597.8 572.4 25.4 75.9 72.6 4.2 5.6 Edmonton 751.0 536.4 512.4 24.1 71.4 68.2 4.5 5.0 All CMAs total 15,769.4 10,732.6 10,041.4 691.2 68.1 93.6 6.4 100.0 M-T-V 8,485.5 5,746.5 5,363.9 382.6 67.7 93.3 6.7 53.5 Rest of CMAs 7,283.9 4,986.1 4,677.5 308.6 68.5 93.8 6.2 46.5 Labour force statistics are designed to represent all persons in the population 15 years of age and over residing in the provinces of Canada, with the exception of persons living on Indian reserves, full-time members of the armed forces, and people living in institutions. Adapted from: Statistics Canada (2002) Labour Force Information for the Week Ended July 20, 2002. (Table 14) Catalogue no. 71-001-PIB. 61 As in the case of Montreal, ethnic and cultural diversity are the norm throughout much of city (Anisef and Lamphier, 2003; Lemon, 1996; Olson and Kobayashi, 1993). Toronto also has the largest number and the largest share of immigrants in Canada (Table 1). While it receives more refugees than Vancouver (Chiu, 2003; Mendez, 2006), Toronto is believed to benefit from the secondary migration of economic immigrants (skilled workers and business class) from Montreal and other places (Ley and Smith, 2000). Immigration is a major contributor to population growth in the city. Heisz (2005: 6) reports that Toronto's population grew by 55 percent (1.6 million people) between 1981 and 2001, and close to two-thirds of that growth was explained by the flow of immigrants that arrived during that time. This massive population growth has unavoidably created a large demand for housing, especially in the suburbs—and beyond. Heisz notes that "in 2001, nearly 10% of the population of Oshawa [a neighbouring C M A ] lived in Toronto five years earlier," and it is believed that many of these new Oshawa residents continued to work in Toronto. Vancouver CMA Vancouver C M A has the third largest population in Canada. It ranks third in population density, behind Montreal and Toronto at 690.3 individuals per square kilometre in 2001. Vancouver is also third in terms of primary destinations for immigrants to the country. One thing where Vancouver outranks Toronto and Montreal is in its housing prices: it is by far the most expensive market in Canada (Figure 2), and has been so for many years despite the deep recession of the 1980s (Barnes et al, 1992; Hiebert, 1999) and the sluggish progress of the average family income. But a stratospheric price-to-income ratio did not prevent the city's homeownership rate from 62 being only slightly lower than Toronto's in 2001 (Figure 8). As we will see in the next chapter, though, owner occupancy in Vancouver does not equate with high income for the majority of households. There seems to be here a strong preference for buying rather than renting one's housing, in spite of the high average dwelling prices and relatively low average incomes that prevail. This suggests that a "distinctive culture of property" based on homeownership—not unlike the culture of rental occupancy that Ray (1999) identified in Montreal—may have become a feature of this C M A . As in the case of Toronto, the single-family detached unit was the most important structural type of dwelling in the city-region in 2001, followed by apartments in small buildings and a small proportion of flats in high-rises (Figure 9), increasingly common along the downtown waterfronts. Residential development in and around downtown Vancouver has dramatically altered the central city's skyline, and densification has become the trademark of the inner core. Nonetheless, the metropolitan area has been experiencing a trend toward suburbanisation similar to Toronto and Montreal. According to Heisz (2005), a worker employed in downtown Vancouver earned on average 10 percent more than a worker employed elsewhere in the C M A . With better incomes, downtown workers are able to bid up residential prices in the area, pushing lower income households out towards some of the adjacent municipalities and beyond, where housing prices are often lower. As in the case of Oshawa and Toronto, Vancouver's neighbouring C M A of Abbotsford has also attracted a considerable number of Vancouver workers: Heisz (2005) reports that one third of Abbotsford's workforce commutes an average of 35 kilometres to work in Vancouver daily each way. However, Vancouver was the only C M A among the Montreal-Toronto-Vancouver trio to post significant population growth 63 in the city centre between 1991 and 2001—almost two-thirds as high as that which took place five kilometres or more from it. Vancouver's population growth also stems in large part from the settlement of newcomers: in fact, between 1986 and 2001, immigration contributed on average a stunning 1.4 percent annually to the city-region's population growth (Heisz, 2005), accounting for 90 percent of all growth between 1991 and 1996 (Hiebert, 1999). Immigrants to Vancouver, Toronto and Montreal arrive from all regions of the world, although Vancouver's location as a "gateway to the Pacific" seems well reflected in its larger proportion of East and Southeast Asian newcomers. Moreover, Vancouver receives a much larger share of business class immigrants than its two metropolitan counterparts, and proportionally fewer family class immigrants and refugees (Ley and Smith, 2000; Chui, 2003). While many types of demographic and socio-economic variation exist across Canadian metropolitan areas, similarities do exist as well. They do not always make material for good news, though. TD Economics (2003: 5-6) observes that: almost every community is grappling with the need for affordable housing, and especially on the rental side. ... [R]oughly one in five or six renter households across Canada is experiencing a severe affordability problem - i.e., paying 50 per cent or more of their income for shelter. Similarly, Canada's largest metropolitan areas have all seen a decline in real wages and growing income disparity in the last 20 years (TD Economics, 2003: 6; Walks, 2001). And yet, in the midst of worsening conditions of social inequality and class polarisation, a price boom is voraciously taking place in the housing markets of Canada's three largest CMAs. 64 In this necessarily brief overview of my three research sites, I have attempted to show that dynamism and structural change are among the most important features shared by Canada's largest cities. This commonality, however, does not translate into sameness with regards to economic conditions, population characteristics, tenure mixes, and residential landscapes. Each metropolitan area is in many ways a distinct 'world in a city' where a variety of social, economic and demographic factors interact to yield variegated inter-metropolitan housing outcomes for different groups of households. In the following chapter, I examine the particular combinations of homeownership entry capacities that existed in each city in 2001, and seek to give them numeric expression in terms of resident tenant households with specific socio-demographic characteristics. 65 C H A P T E R 3: A P R O F I L E O F T E N A N T H O U S E H O L D S A N D T H E I R C A P A C I T Y T O E N T E R H O M E O W N E R S H I P In the previous two chapters, I have sought to provide an overview of the important role that ideological, methodological, socio-economic, historical, and geographical contexts play in shaping housing markets and the conditions of accessibility confronted by local households. This chapter finally turns to the households themselves, examining the uneven possibilities of tenants to transition into homeownership. As the following analysis will show, such variability is not only linked to the metropolitan geographies I identified earlier, but also to the demographic and economic characteristics of a household (narrowly equated here with the characteristics of the person defined in the Census data as the primary household maintainer or 'Person 1'). It is worth keeping in mind that the demographic categorisations to which I now turn—in conversation with a vast academic and policy-related literature on homeownership—are pre-defined by the sources of information to which I have access. But my use of ready-made population groupings does not imply uncritical acceptance of the authority of their descriptions. I am aware that as social constructions, categories thrive on their power to conjure up or alter collective identities and to obscure commonalities and differences that are not recognised by customary social norms (Hacking, 1983). It is precisely because of this defining hold on our everyday understanding of the world that I am interested in the stories that Census classifications can tell us. Statistical categories can help us see beyond our current assessments of what constitute matters of concern (Latour, 2004) by helping us identify the biases and contingencies that underlie such judgements. Despite the historical recurrence of housing 66 access as a crucial social issue in modern societies, nothing is necessarily fixed when it comes to contemporary conceptualisations of the relationships between human beings, their practices of shelter, and the larger social contexts in which they take place. As I explained earlier, my study is limited to tenant households in Canada's three largest metropolitan areas: Montreal, Toronto and Vancouver. Each one of these city-regions has its own geographies of tenure and, consequently, its own spatial particularities concerning the possibilities to transfer from one form of tenure to another, be it for example from owner occupancy to tenancy, or vice-versa, as in the case of the present study. I have already mentioned many of the reasons why such a transition does or does not occur. I will then turn to my own examination of the geographically circumscribed relationship between household income and average house prices. There is an intervening factor acting as mediator in this relationship, namely the market for capital, in the form of mortgage lending institutions with general criteria for the evaluation of their credit allocation process. Table 7 summarises the results of my assumptions on the potential convergence of these three factors. Separating tenant households from homeowner ones, I identify a mortgage-qualifying income threshold and the proportion of tenant households in each C M A that would have made the chartered banking institutions' cut in 2001, based on Census-declared annual before-tax income for the year 2000. This table shows that, all other things held constant, an appreciable proportion of homeowner households would have been unable to qualify for a mortgage on the average-priced house or a moderately priced one had they to lost their housing and any accumulated equity, for instance as a result of foreclosure. (Ideally, one could compare these findings with data on homeowner households spending 50 percent or more 67 Table 7: Tenure distribution of households by mortgage qualifying income thresholds, 2001 Owner Tenant households households Number Column % Number Column. % House price: Yearly pay-ments: Qualifying household income: Montreal: N= 709,850 100 708,450 100 Average: Qualifying income, Group 1 ($45,000 and over) 188,900 26.6 482,350 68.1 $128,550 $13,900 $46,350 Qualifying income, Group 2 ($35,000-$44,999) 88,800 12.5 74,000 10.4 Moderate: Non-qualifying income, Group 3 ( < $35,000) 432,200 60.9 152,050 21.5 $96,400 $11,300 $37,650 Toronto: N= 606,100 100 1,037,900 100 Average: Qualifying income, Group 1 ($75,000 and over) 93,950 15.5 522,700 50.4 $251,500 $23,800 $78,350 Qualifying income, Group 2 ($60,000-$74,999) 58,000 9.6 137,650 13.3 Moderate: Non-qualifying income, Group 3 ( < $60,000) 454,150 74.9 377,550 36.4 $188,650 $18,750 $62,500 Vancouver: N= 295,700 • 100 462,600 100 Average: Qualifying income, Group 1 ($85,000 and over) 28,550 9.7 150,250 32.5 $285,900 $26,500 $88,350 Qualifying income, Group 2 ($70,000-$84,999) 18,750 6.3 54,150 11.7 Moderate: Non-qualifying income, Group 3 ( < $70,000) 248,400 84.0 258,200 55.8 $214,450 $20,800 $69,350 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Sources: C M H C (2005); 2001 Survey of Household Spending (Table 62F0035); Bank of Canada (CANSIM Table V122521). of their income on housing, to obtain an assessment of how many home-owning households are at risk of losing their homes.) However, Table 7 also suggests that for at least some households, a low income-to-house price ratio is not a good predictor of the ability to transition into homeownership (see for example Hulchanski, 1995; 2005)— particularly in Vancouver, where a stunning 55.8 percent of home-owning households have incomes below the qualifying threshold for a mortgage on a moderately priced 68 dwelling. This is an important point to keep in mind, and one to which I will return in the concluding chapter. We can now begin to form an empirically grounded sense of the difference between the accessibility measures I advocate in this thesis and the affordability measures published by C M H C and Statistics Canada (namely the Core Housing Need measure and INALH). In the three cities covered by my study, the average income of households considered to be in core housing need in 2001 was considerably lower than the minimum mortgage-qualifying income according the assumptions I laid out earlier (Table 8). This difference in results between the two measures suggests that tenant households wishing to exit rental housing in favour of owner occupancy face a much more imposing financial barrier than in the case of accessing what C M H C defines as acceptable rental housing. This is especially so in Vancouver, where the average-income-to-average-house-price ratio is by far the highest of the three CMAs (Figure 4). But the primary concern of this thesis is not the comparison of affordability measures, but the outcomes of households lacking owner status and, presumably, house-Table 8: Comparison of homeownership accessibility income thresholds and average incomes of tenant households in core housing need, 2001 Minimum mortgage-qualifying hhld. income Avg. income of tenant hhlds. in CHN* Average STIR* Montreal $37,650 $13,178 50% Toronto $62,500 $22,995 4 5 % Vancouver $69,350 $18,693 49% Note: The minimum mortgage-qualifying household income in this table refers to the moderately priced dwelling as defined in this study. * C H N stands for Core Housing Need; STIR is the acronym for 'shelter-cost-to-income-ratio.' Sources: C M H C (2005); 2001 Survey of Household Spending (Table 62F0035); Bank of Canada (CANSIM Table V122521). 69 related equity. Returning to Table 7, we see that tenant households in Vancouver were clearly worse off than in the other two CMAs under analysis, as 84 percent (248,650 households) were unable to access even the moderate mortgage. In Toronto, 74.9 percent of tenant households (454,150 in total) were classified into Group 3, and in Montreal, 60.9 percent (432,200 households) were in the same financial situation relative to their own C M A ' s housing market. The three right-most columns in table suggest that these proportions are at least partly associated with house prices in each city, as Montreal's average figure is, at $128,550, about half of Toronto's $251,500, and more than 50 percent below the $285,900 average price in Vancouver. Table 7 shows therefore that to a large extent, the ways in which house prices and current mortgage-borrowing conditions separate tenant households along income lines are locally contingent. In all three CMAs, more than 50 percent of tenant households had incomes that would not have qualified them for even a moderately priced dwelling as defined in this study. The proportion of tenant households in this financial situation was in fact almost three quarters of the total in Toronto, and a stunning 82.1 percent in Vancouver. Already in 2001, the 2000 median income in these two cities ($59,502 and $49,940 respectively, as per Statistics Canada CAT. No.97F0020XCB 2001005, figures not shown) was not enough to qualify for a moderate mortgage. Given that both C M A s had household homeownership rates of more than 50 percent, one can say. that an appreciable number of owner-occupied dwellings were inhabited by households with incomes at or near the 2000 median. Indeed, 36.4 percent of home-owning households in Toronto and 55.8 percent in Vancouver would not have had sufficient reported incomes to qualify for a moderate mortgage in 2001 had they lost their equity somehow (Table 7). 70 In Montreal, where the 2000 median income of $42,123 would have qualified a household for a mortgage on a moderately priced house, the proportion of homeowner households with below-qualifying incomes was much lower but still considerable, at 21.5 percent. I cannot elaborate on the meanings, and implications of these figures here, considering that not much is known about these 'house-rich-and-cash-poor' households and the financial sustainability of their owner occupancy status. But these statistics clearly demonstrate the need for Canadian research that replicates and expands on U.S. and U.K. studies of the homeownership experiences of low-income households (e.g., Reid, 2004; Ford et al, 2001). Differences in income by place of residence are not the only source of heterogeneity among tenant households. The remainder of this chapter examines a series of demographic variables that were found to act as further markers of differentiation in the capacity of households to enter homeownership. Most of these household characteristics have already been identified in the literature as determinants of variability in tenure outcomes (Edmonston, 2004; Haan, 2005a; Lapointe Consulting Inc. and Murdie, 1996; Laryea, 1999). Some of these variables apply to the tenant population at-large, while others focus specifically on immigrant-headed households. As discussed earlier, a series of benefits at both the individual and social level are commonly attributed to homeownership; at the same time, the unevenness of its accessibility has been associated with certain negative effects for individuals and society, including a broadening of wealth differentials between tenure classes and the potential transmission of such disparities from one generation to the next. In order to form an understanding of the socio-demographic factors that play a role in the development of 71 this wealth gap effect, the following analysis explores the characteristics of non-qualifying tenant households relative to the rest of the household universe in Montreal, Toronto and Vancouver. Holding everything else constant, what was the likelihood that households exhibiting a given socio-demographic feature were already priced out of the homeownership market in 2001, according to the conditions and assumptions laid out in Chapter 2? It is important to recognise that the Census cross-tabulations that I have produced in an attempt to explore answers to this question represent the 'stock' of tenants who would (or, conversely, would not) have qualified for a moderate of average mortgage in 2001 (Listokin et al., 2002). As aggregate snapshots, these figures cannot be equated with the actual 'flows' of tenancy transition that took place that year. These cross-sectional results are similarly unable to account for the dynamic nature of households and their Primary Maintainers: marriages occur, degrees are obtained, children are born, and so on, transforming the characteristics of households from one moment to the next. A set of graphs, included in the Appendix, accompanies the tables presented this chapter. Age In housing studies, the notion of'life ladder' or 'life cycle' (Bourne, 1981; Foote et al, 1960, cited in Haan, 2005a; Murdie et al, 1999; Perin, 1977, cited in Doucet and Weaver, 1991) has been employed to describe patterns of evolving housing needs and the changing financial ability to satisfy them at various stages throughout the lifespan of the average household. Historical observation of aggregate household trajectories from early adulthood to old age seems to support this notion. Tracing the relationship between 72 homeownership and age in Hamilton, Ontario from 1872 to 1981, Doucet and Weaver (1991: 311) found that, despite variations attributable to changing economic conditions through time, "[youth] invariably stood more precariously on a lower rung of home ownership than seniors, with the middle-aged groups attaining the intermediate rungs. In most of the sample years, the over-64 age group had a higher percentage of homeowners than any other age group...." While this pattern was fairly constant for most of the surveyed years (controlling for periods of economic downturn), the proportion of homeowners in the latter age group began to show a decline relative to younger cohorts by the 1960s, and continued to fall until 1981, the last year of the study. The 2001 age-group tenure distribution in the three CMAs covered by my study is consistent with the city of Hamilton's historical pattern of homeownership attainment (Table 9), which in fact seems to extend also to the 2001 pattern of homeownership accessibility. In the case of the youngest age cohort, low homeownership rates and high proportions of non-qualifying households are the norm in the three cities. These results were expected, due to the typically lower wages earned by individuals in the first few years after leaving the parental household. Montreal stands out, however, for its high rate of tenant households who would qualify for an average mortgage relative to Toronto and Vancouver, a feat that is repeated by the 25 to 34 cohort. That these households were holding off on home ownership despite their relative income capacity may be explained for example by a lack of savings to afford a down payment, a desire to wait until a larger and better located dwelling is within their means, an expectation of little or negative housing price appreciation, a reluctance to take on a 'grown-up's' commitment, or by the 73 Table 9: Qualifying tenant households, by age of Person 1 ("Primary household maintainer") (Number of households and proportions of all tenants. Column percentages, 2001) 80 and 15-24 25-34 35-44 45-54 55-64 65-79 over Total Montreal: Total homeowners 5,150 74,700 180,100 179,350 127,350 119,950 21,850 708,450 Total tenants 60,334 154,998 152,497 122,823 78,437 104,819 35,925 709,833 Group 1 16.8 32.7 31.6 32.6 24.8 15.0 13.0 26.6 Group 2 10.6 14.2 15.4 13.7 11.3 8.4 6.3 12.5 Group 3 72.5 53.1 53.0 53.7 64.0 76.6 80.7 60.9 Toronto: Total homeowners 5,750 111,450 260,250 255,400 175,950 183,450 45,600 1,037,900 Total tenants 29,450 155,000 162,350 110,300 61,150 64,350 23,450 606,050 Group 1 7.6 19.0 15.8 18.5 16.7 7.6 4.1 15.5 Group 2 7.3 10.7 10.1 11.5 9.8 5.3 3.0 9.6 Group 3 85.0 70.2 74.1 69.9 73.5 87.2 93.0 74.9 Vancouver: Total homeowners 3,300 44,650 103,700 117,650 81,300 86,350 25,650 462,650 Total tenants 21,750 81,650 75,650 54,800 24,650 26,100 11,150 295,700 Group 1 3.8 11.0 11.6 12.1 10.2 2.3 2.0 9.7 Group 2 3.2 7.4 7.4 7.7 4.8 2.8 2.0 6.3 Group 3 92.9 81.6 80.9 80.2 85.2 94.9 95.9 84.0 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). C M A ' s cultural predisposition to 'lifestyle renting' (Varady and Lipman, 2004) that Brian Ray (1999) has described. In the case of tenant household subgroups in which Person 1 ranges in age between 25 and 64 years (the most economically active years in the labour force), the three C M A s showed rates of inaccessibility below the local average, except for the 55-64 years age group in Montreal, which actually has a higher rate than the C M A as a whole. In fact, the 55-64 years age group generally has higher rates of inaccessibility than other age groups between 25 and 64 years (the prime years of the economically active population) in the three CMAs under study. One possible explanation is that people in this birth cohort entered their prime first-home buying age (30 to 40 years old) in the 74 mid-1970s to mid-1980s, during a period of recession and major labour force restructuring. In Montreal, this age group reached the prime first-home buying age just as dozens of companies and massive amounts of investment capital began their exodus from the city due to mounting Quebecois support for the separatist cause. The ensuing economic downturn may have paired a dropping demand for housing with a dried up supply, as residential construction came to a halt shortly after the election of Rene Levesque in 1976. More than 50,000 Montreal households in this age cohort were not living in owner occupied housing by 2001, and had apparently been unable to catch up to the home buying capacities of younger tenant cohorts. This observation suggests that the unfavourable economic climate that prevailed at a crucial stage in these households' life cycle may have had long lasting consequences for those who did not enter homeownership early on. Further research would be needed to verify this hypothesis. If it were found to be true, one could expect the price boom of the early millennium to have long-term consequences for many households that have been priced out of homeownership. Person l 's age group provides an insight into the level of polarisation that may exist, or could potentially develop in the long run, when households that belong to subgroups with high levels of homeownership are continuously unable to enter homeownership. In Vancouver, households in which Person 1 individuals are aged 65 to 79 had the highest likelihood of having attained owner occupancy in 2001, but the rate of households in Group 1 and Group 2 was minuscule. That is to say, practically all of the PHMs that could afford to buy a house had already done so by the time they reached that age (Figure 10 in the Appendix). While there may have been many other individual and 75 group factors at play, it.is not unreasonable to believe that a considerable number within the 24,750 'remaining' tenants must have faced serious constraints compared to other households in the same age group. Household composition While different age groups seem to navigate their housing careers in accordance with the expectations of the average household's life cycle, the transformation of household compositions in the past two decades is challenging some of the model's traditional precepts. Michael Haan (2005a: 11) finds for example that economic family . structures are changing for Canada as a whole, with larger shares of unattached individuals among the Canadian-born (from 19% of all families in 1981 to 24% in 2001). Haan also reports that an increase in the prevalence of lone parent households from 1981 to 2001 went from a rate of seven percent to 12 percent for immigrants and from nine percent to 10 percent for the Canadian-born. In terms of family-type households, the traditional life cycle model and the consumer choice model would attribute the highest rates of homeownership to married couples with children, while in fact all family types were "closing the homeownership gap with married couples with children" (Haan, 2005a: 25) in 2001. How are these transformations affecting tenant households in their ability to enter homeownership? Table 10 shows that in the three CMAs, non-family one-person' households had the highest likelihood of not qualifying for a mortgage on a moderate or average dwelling, followed by lone-parent families. It is worth noting that persons living alone are the second largest household type in the three CMAs (see Figure 11 in the Appendix). The fact that these smaller tenant households have the lowest capacity of 76 Table 10: Qualifying tenant households, by household composition (Number of households and proportions of all tenants. Column percentages, 2001) Lone parent families Couples, with or without children Multiple family households Non-family households: 1 person only Non-family households: 2 persons or more Total Montreal: Total homeowners 61,000 513,100 10,250 110,850 13,250 708,450 Total tenants 102,488 235,366 5,211 321,936 44,832 709,833 Group 1 20.8 46.6 58.8 12.2 34.6 26.6 Group 2 13.9 15.6 12.1 9.7 13.1 12.5 Group 3 65.2 37.8 29.1 78.1 52.3 60.9 Toronto: Total homeowners 88,150 729,350 53,900 144,450 22,000 1,037,900 Total tenants 91,000 240,200 13,200 217,850 43,830 606,050 Group 1 8.4 23 ' 44.0 5.4 30.5 15.5 Group 2 7.0 13.1 15.1 5.6 13.4 9.6 Group 3 84.6 63.8 40.9 89.0 56.3 74.9 Vancouver: Total homeowners 35,800 301,650 21,100 93,000 11,050 462,650 Total tenants 35,000 107,750 3,700 121,250 27,950 295,700 Group 1 4.7 17.9 16.0 2.2 15.3 9.7 Group 2 4.9 10.2 17.9 2.2 9.7 6.3 Group 3 90.4 71.9 65.9 95.6 75.0 84.0 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). entry while multiple family households (namely households made up of families living with another family, as in the case of young married couples who live with in-laws) have the highest likelihood of meeting the mortgage-qualifying threshold (as well as the highest rates of homeownership) in the three cities suggests that larger households may have more members who are in the labour force than just the P H M and his or her spouse. Obviously, it would be expected that additional contributors to household income would have a dramatically positive effect on a household's capacity of entry into homeownership. Ley (1999) has suggested that additional earners could explain why immigrants on average have higher household incomes than the Canadian-born, who on average have smaller households. Further research is needed to verify these hypotheses 77 and help us understand the mechanics of how household structure affects the financial accessibility of homeownership. Household size (as a separate Census variable from household structure) will be discussed further below. Presence of children Children may provide a family with an additional incentive to seek the security of tenure associated with owner-occupied housing. But the costs associated with raising children in metropolitan settings may discourage families from having children in the first place. Haan (2005a: 11) reports that the proportion of Canadian-born PHMs in married families with children has declined from 40 percent in 1981 to 32 percent in 2001. The same high costs may also act as a barrier to enter homeownership for tenant families who become parents before they are able to purchase their own home. The relationship between presence of children and the household's ability to access homeownership in the three C M A s in this study is in fact geographically contingent. Couples with children were more likely than couples without children to meet the mortgage-qualifying income thresholds laid out in this research for moderately and average priced dwellings in Montreal, but the reverse was true in Toronto and Vancouver (Table 11). In all three CMAs, though, homeownership was more likely for couples with children than for those without. This discrepancy between ownership rates and mortgage-qualifying rates for couples with and without children may be related to the type of childcare programme available in each metropolitan area. In 2001, childcare was more deeply subsidised than in Vancouver and Toronto. Lower childcare costs not only would make parenthood financially more feasible for tenant households in Montreal, but it would have allowed both parents to have access to the workforce, increasing the total 78 Table 11: Qualifying tenant households, by presence of children in the household (Number of households and proportions of all tenants. Column percentages, 2001) Multiple-Lone-parent families Couples, with children family house-holds, with children Couples, no children All other households, no children Total Montreal: Total homeowners 59,000 313,100 5,900 198,200 128,000 708,450 Total tenants 102,500 114,950 3,150 , 120,400 368,850 709,850 Group 1 20.8 48.4 55.1 44.9 15.2 26.6 Group 2 13.9 16.5 13.0 14.7 10.1 12.5 Group 3 65.2 35.1 31.7 40.4 74.6 60.9 Toronto: Total homeowners 84,050 488,250 37,950 234,450 177,800 1,037,900 Total tenants 91,000 144,400 9,700 95,800 265,150 606,050 Group 1 8.4 21.0 41.1 26.1 10.1 15.4 Group 2 7.0 13.1 16.3 13.1 7.0 9.6 Group 3 84.6 65.8 42.6 60.8 82.9 74.9 Vancouver: Total homeowners 34,650 179,900 14,450 118,850 108,450 462,650 Total tenants 31,650 42,900 2,900 49,750 137,400 295,700 Group 1 4.7 17.2 16.4 E 18.8 4.7 9.6 Group 2 4.9 8.8 19.0 E 11.7 3.7 6.3 Group 3 90.4 74.0 64.6 E 69.5 91.6 84.0 F: Too unreliable to be published. E : Sample size for this category is small. Use this figure with caution. Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). amount of income earned by the household. In all three CMAs, however, all other types of household without children were, along with lone parent households, the least likely to qualify for a moderate or average mortgage when considered as an aggregate group. While this result somehow confounds the childcare hypothesis, Table 11 opens the door for further research to establish whether the lack of subsidised childcare may have an interaction effect with other variables and indirectly affecting the financial accessibility of homeownership. 79 Marital status Homeownership rates have traditionally been higher for married couples than for common-law couples and other family and non-family household types, although a considerable amount of variation existed across metropolitan areas (Edmonston, 2004; Laryea, 1999). But since 1996, more and more common-law couples have been buying homes than was previously the case (Thomas, 2005: 3.8). This not only alters traditional expectations of the life-cycle and consumer choice models with regards to homeownership, but as Table 12 shows, it also affects expectations about the financial capacity of tenant households to transition into owner occupancy. According to the traditional model, one would expect married couples with children to have the highest rates of homeownership, followed by common-law couples with children, married couples without children, and common-law couples without children. The expectation for tenant households to qualify for the moderate and average mortgage would follow a similar order. But the 2001 Census results disprove these expectations. Married couples with or without children who had not already purchased their home by the time of the 2001 Census were less likely to qualify for the moderate and average mortgage than common-law couples with or without children in the three CMAs. In Toronto and Vancouver, married couples with children were slightly less likely than married couples without children to qualify for the moderate or average mortgage defined in this study. Regardless of marital status, couples in single-family households were overall considerably more likely than 'Other households' to meet the mortgage-qualifying income thresholds and to be homeowners as well, except in Vancouver where common-law couples without children were less likely to be homeowners than households in the 80 Table 12: Qualifying tenant households, by marital status of Person 1 (Number of households and proportions of all tenants. Column percentages, 2001) Married couples, no children Married couples, with children Common-law couples, no children Common-law couples, w/ children All other house-holds Total Montreal: Total homeowners 144,600 244,200 54,350 69,950 195,350 708,450 Total tenants 63,350 79,950 57,100 35,000 474,450 709,850 Group 1 35.4 46.9 55.4 51.7 16.7 26.6 Group 2 15.1 16.6 14.3 16.3 11.0 12.5 Group 3 49.5 36.4 30.3 32.1 72.3 60.9 Toronto: Total homeowners 209,500 474,700 27,450 17,650 308,550 1,037,900 Total tenants 65,700 130,600 30,100 13,800 365,850 606,050 Group 1 22.6 20.2 33.8 29.2 10.5 15.5 Group 2 11.6 12.8 16.4 16.6 7.2 9.6 Group 3 65.7 67.1 49.8 54.2 82.2 74.9 Vancouver: Total homeowners 104,400 174,600 15,550 7,150 160,950 462,650 Total tenants 30,750 51,000 19,050 7,050 187,950 295,700 Group 1 17.1 17.2 21.6 17.3 4.9 9.7 Group 2 10.6 8.4 13.6 12.1 4.1 6.3 Group 3 72.3 74.4 64.7 70.3 90.9 84.0 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). "Other" category. One possible explanation is that more couples may be living common law as tenants, before wedlock, but switch to owner occupancy upon marriage, leaving in only the most impoverished married couples (and those saving for an above-average first-home) in rental housing. Marital status is not the first variable in this study to exhibit discrepancies between the homeownership rates of a sub-category and the financial ability of tenant households in that same sub-category to access owner occupancy (age group and presence of children have provided examples of a similar kind of inconsistency). It seems safe to observe at this point that homeownership attainment is not necessarily a good 81 predictor of the incidence of financial inability to access owner occupancy among tenant households. Household size Large dwellings that can accommodate larger households may be more expensive to buy, but larger households can potentially have more members contributing to household income, as Ley (1999) has suggested in the case of immigrant households. Haan's (2005a:20) 1981-2001 cross-tabulations reveal that "income and the number of full-time earners in a family increases the probability of homeownership by a large margin," which helps explain why larger households typically have higher rates of homeownership than smaller ones, although some degree of variation can be expected across CMAs. Table 13 indicates that the share of tenant households that would have been unable to qualify for the moderate and average mortgage in this study tends to grow in tandem with household size. In the three C M A s under analysis, tenant households composed of persons living alone have the highest likelihood of not qualifying, with rates ranging from 78.1 percent in Montreal to an astounding 95.6 percent in Vancouver. Two-person tenant households had the second largest share of non-qualifying households in Montreal and Vancouver, but occupied the third spot in Toronto. Three-person households were the second largest group there, but only by a small margin. These figures are of concern in the short run, first because one-person and two-person households already have the lowest rates of homeownership, but secondly because together they make up the largest share of total households in all three CMAs. It would appear that at least for these two categories of households, size could be acting as an 82 Table 13: Qualifying tenant households, by household size (Number of households and proportions of all tenants. Column percentages, 2001) Two Three Four Five Six persons One person persons persons persons persons or more Total Montreal: Total homeowners 110,850 236,600 140,400 149,200 51,700 19,700 708,450 Total tenants 321,950 212,350 91,600 53,500 20,300 10,150 709,850 Group 1 12.2 35.1 40.5 46.2 42.4 45.4 26.6 Group 2 9.7 14.2 15.3 15.7 16.9 14.6 12.5 Group 3 78.1 50.7 44.2 38.1 40.6 40.1 60.9 Toronto: Total homeowners 144,450 282,300 194,650 242,500 108,400 65,600 1,037,900 Total tenants 217,850 166,450 95,600 72,450 34,300 19,400 606,050 Group 1 5.4 20.2 19.0 22.0 23.2 32.8 15.5 Group 2 5.6 11.6 11.6 11.8 13.1 11.8 9.6 Group 3 89.0 68.2 69.4 66.2 63.7 55.3 74.9 Vancouver: Total homeowners 93,000 139,950 76,500 86,300 39,750 27,150 462,650 Total tenants 121,250 87,450 40,350 28,350 11,450 6,850 295,700 Group 1 2.2 13.3 15.5 16.0 17.4 21.5 9.7 Group 2 2.2 9.4 8.2 7.9 10.6 15.1 6.3 Group 3 95.6 77.4 76.3 76.0 71.9 63.1 84.0 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). important determinant of the ability of tenant households to transition into owner occupancy. There are anecdotal reports that the mortgage-finance and residential construction industries have started to respond to this phenomenon by developing novel products tailored specifically to the needs of small households, such as "interest -only" mortgages that initially reduce the household's regular outlays to the financial institution, and studio-type condominium apartments meant to provide small households the opportunity to get a foot in on into the housing ladder. The expectation is that over time, many of these households will grow in both size and income. Schooling 83 Education is generally regarded as a good predictor of an individual's financial . and social capital in the present and the future, which in turn is expected to affect the homeownership entry capacity of PHMs. One would therefore expect that higher levels of schooling would improve the capacity of tenant households to qualify for moderate and average mortgages. However, Ley and Smith (1997) have reported a weak correlation between failure to complete high school and a family's incidence of poverty as measured by Statistics Canada's Low Income Cut-Off (LICO). The distribution of the household universe across the three C M A s in this study shows a large number of households headed by individuals who have not completed high school (Figure 15 in the Appendix). In Vancouver, the ownership rates of these households are higher than those of households headed by individuals with some post-secondary and with a bachelor's degree or above (Table 14, percentages not provided). Moreover, only in Montreal is there a lower likelihood of homeownership attainment for households where Person 1 has not completed high school in relation to all other households. (These results are consistent with Laryea's (1999) findings using 1991 Census data.) However, due to potential interaction effects related to the age composition of these two educational attainment groups, further disaggregation by Person-1 age group may shed some light on the reasons behind these counter-intuitive and uneven 'outcomes. In terms of tenant households who would not have qualified for the moderate and average mortgages posited in this study, the highest share is among households where the P H M has not completed high school, and lowest among households where PHMs have obtained at least a bachelor's degree (Table 14). In Montreal, more than 50 percent of 84 Table 14: Qualifying tenant households, by highest level of schooling of Person 1 (Number of households and proportions of all tenants. Column percentages, 2001) High Trades Bachelor Less than school certificate/ Some post- and high school grad diploma secondary above Total Montreal: Total homeowners 169,150 109,350 30,700 237,500 161,750 708,450 Total tenants 220,900 104,050 21,600 235,350 127,950 709,850 Group 1 14.9 26.0 28.7 28.1 44.2 26.6 Group 2 9.2 12.3 14.0 15.4 12.8 12.5 Group 3 75.9 61.7 57.3 56.5 43.0 60.9 Toronto: Total homeowners 244,800 115,100 38,600 356,800 282,600 1,037,900 Total tenants 147,950 67,900 15,250 217,550 157,450 606,050 Group 1 8.9 12.7 16.5 14.9 23.6 15.5 Group 2 6.0 9.3 8.7 10.1 12.4 'l 9.6 Group 3 85.1 78.1 74.8 75.0 64.0 74.9 Vancouver: Total homeowners 101,150 50,050 15,900 185,350 110,200 462,650 Total tenants 61,650 28,450 6,800 126,650 72,150 295,700 Group 1 4.1 7.7 6.0E 9.1 16.5 9.7 Group 2 3.1 7.8 7.1E 6.5 8.1 6.3 Group 3 92.8 84.6 87.0 84.4 75.3 84.0 E : Sample size for this category is small. Use this figure with caution. Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). tenant households in this last group had sufficient income to qualify for a mortgage on a moderately priced house, and almost 40 percent could afford an average mortgage. Tenant households in all other subcategories had rates of non-qualifying households more or less at par with the C M A s ' aggregate totals. One arresting feature of C M A tenure distributions by highest level of schooling is that households where the P H M held a trades certificate or diploma had the highest rates of homeownership in the three CMAs. Not all trades people fared as well (witness the numbers of trade-diploma/certificate tenant households in Group 3) but the small share of the overall population with this type of training might imply that there is less competition 85 in some fields, and therefore higher incomes for their practitioners. In Vancouver, a potentially polarising situation exists within this predominantly homeowner subcategory, as the proportion of all households that are tenants in Group 1 and 2 is very small. It is also curious to note that in the three CMAs, homeownership rates were very similar between PHMs who had as a maximum a high school diploma and PHMs who had some post-secondary education below a bachelor's degree. In Vancouver, PHMs with a bachelor's (but not higher) degree were considerably less likely (63.8 percent to 58.7 percent, figures not shown) than those with a maximum of a high school diploma to live in owner occupied dwellings, while in Montreal and Toronto, having a bachelor's degree but not higher yielded only a moderately improvement on the probability of living in owner occupied housing (figures not shown), confirming Ley and Smith's (1997) findings. Again, further disaggregation could help assess the extent to which an age interaction effect might help explain such outcomes. There is another possible reason why there are relatively high rates of Group 1 and Group 2 tenants who have not made the transition to owner occupancy in the case of households where the P H M has a bachelor's degree (or higher). These households may already be carrying a high debt load as a result of student loans, which can amount to tens of thousands of dollars for graduates with advanced degrees. This presents another potential accessibility barrier that is worth exploring further in a separate research project. Language Ley and Smith (1997: 45) have also explored the effect of the PHM's use of the C M A ' s official language on the incidence of low income among economic families. 86 Table 15: Qualifying tenant households, by primary language spoken at home by Person 1 (Number of households and proportions of all tenants. Column percentages, 2001) Official language Non-official language*, but knows E/F Non-official language*, and doesn't know E/F Total Montreal: Total homeowners 636,000 65,050 7,400 708,450 Total tenants 620,600 80,050 9,000 709,700 Group 1 27.2 23.1 18.0 26.6 Group 2 12.7 11.2 7.8 E 12.5 Group 3 60.1 65.7 74.2 60.9 Toronto: Total homeowners 775,500 226,500 35,850 1,037,850 Total tenants 432,000 151,950 22,000 606,000 Group 1 17.2 11.7 7.4 15.5 Group 2 9.9 9.2 4.7 9.6 Group 3 72.8 79.0 87.9 74.9 Vancouver: Total homeowners 357,100 86,850 18,700 462,650 Total tenants 234,250 50,850 10,400 295,550 Group 1 10.8 5.7 F 9.7 Group 2 7.0 4.2 F 6.3 Group 3 82.1 90.2 95.9 84.0 * Aboriginal languages not included in these figures. F: Too unreliable to be published. E : Sample size for this category is small. Use this figure with caution. Totals may not add up due to non-response and rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). Using Census data from 1991, they found "consistently positive, but modest, correlations," both in terms of the language used at home and of the ability to speak English or French. By contrast, Haan (2005a: 22) finds that between 1981 and 2001, knowledge of the official language is a negative predictor of homeownership, which "may have to do with the high homeownership rates of some 1960s arrivals that do not know English or French." The evidence provided in Table 15 lies somewhere between these two positions. Households in which the P H M had no knowledge of an official language had the highest 87 rates of non-qualifying tenant households as defined in this study, with rates well above the metropolitan totals. In fact, households in which Person 1 spoke an official language at home were only slightly more likely to have incomes above the mortgage-qualifying thresholds for moderately and average priced dwellings in all three CMAs. Language therefore seems to be a somewhat more decisive factor for immigrant households that have not yet entered owner occupancy than it is for those who arrived with the means to purchase a home early on. Further evidence of this hypothesis is provided by the tenure distribution of households as a function of the Census language variables (Table 15, percentages not provided). Households in which the P H M speaks an official language at home were more likely to own their housing in Montreal and Toronto, but were slightly less so in Vancouver, where business class immigrants have been found to exert an appreciable effect on the housing market (Ley, 1999; 2001). What we see here, then, is perhaps not so much evidence of the explanatory power of the language variable, but rather an effect of the appreciable number of immigrants in Vancouver C M A who have been able to enter homeownership from a very early stage in the settlement process (Hiebert et al., Forthcoming; Mendez et al, 2006). Vancouver once again provides a potential instance of polarisation in the making, as households where the P H M did not speak an official language at all recorded a high likelihood of being homeowners, yet also had extremely low rates of tenant households in Groups 1 and 2 (Figure 16 in the Appendix). 88 Citizenship The citizenship variable functions as a sort of bridge between the two types of variables that I am including in this analysis, namely those that represent socio-demographic features that change or can be changed over time and those that represent characteristics that normally do not change over the lifetime of a household or P H M . An immigrant can apply for Canadian citizenship i f he or she has lived in Canada for at least three of the four years that preceded submission of an application, provided he or she fulfills certain conditions. For immigrants, therefore, citizenship is a variable that fits within the first set of variables. For Canadian citizens by birth, continued residence in Canada is normally not associated with renouncing their nationality; for them, Canadian citizenship functions like an immutable characteristic, and for this reason it would seem to fit better with the second set of variables in this study. Tenant households in which the P H M was a non-permanent resident or a non-naturalised immigrant with 10 years or less in Canada had the highest probability of not qualifying for a mortgage on either a moderate or average priced house in the three C M A s (Table 16). The latter were considerably more likely not to meet the qualifying income threshold than naturalised immigrants who had been in Canada for a similar period of time. Canadian PHMs by birth had the lowest likelihood of living in a household that would not have qualified for the moderate or average mortgage in the three CMAs, although in Montreal the likelihood is only marginally lower than for all tenant households in the C M A as a whole. This result gives the impression that a positive association exists between naturalisation and the ability to access homeownership in the case of immigrants with 89 Table 16: Qualifying tenant households, by citizenship and period of immigration of Person 1 (Number of households and proportions of all tenants. Column percentages, 2001) Canadian by Canadian by birth naturalisation Other citizenship Total, Immigrated before 1991 Immigrated between 1991 &2001 Immigrated before 1991 Immigrated between 1991 &2001 Non-permanent residents Montreal: Total homeowners 565,350 119,350 9,550 8,800 4,200 1,300 708,450 Total tenants 538,800 85,650 31,800 8,550 32,050 12,950 709,850 Group 1 27.7 24.7 23.8 28.9 18.1 19.4 26.6 Group 2 12.8 12.6 14.2 10.0 9.8 6 .0 E 12.5 Group 3 59.5 62.8 62.0 61.4 72.0 74.6 60.9 Toronto: Total homeowners 481,850 408,400 62,500 45,600 36,200 3,350 1,037,900 Total tenants 270,400 150,700 69,000 22,900 79,450 13,600 606,050 Group 1 19.2 14.1 13.2 13.7 8.6 13.6 15.5 Group 2 10.5 9.4 9.5 8.1 8.2 3.5 9.6 Group 3 70.4 76.5 77.3 78.3 83.2 82.7 74.9 Vancouver: Total homeowners 259,600 135,000 33,350 14,000 19,400 1,300 462,650 Total tenants 181,050 50,750 21,050 7,850 26,350 8,650 295,700 Group 1 11.1 8.4 7.5 9.4 E 5.0 7.3 E 9.7 Group 2 7.0 6.2 6.1 7.5 3 .1 E 3.0 E 6.3 Group 3 82.0 85.4 86.3 83.3 91.8 89.4 84.0 E : Sample size for this category is small. Use this figure with caution. Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). stays of 10 years or less in Canada, but not so much in the case of immigrants who have had longer stays. However, it is also quite possible that an additional factor or set of factors could be driving both the preference for naturalisation and the ability to enter owner occupancy, including for example the employment situation of the members of the household. Further research would be needed to establish whether this apparent relationship is statistically robust and non-spurious. 90 Sex As I discussed in the previous chapter, the use of primary household maintainer data as proxies for the characteristics of a household is a problematic solution to the challenge posed by the statistical need for simplification in housing consumption studies. Doucet and Weaver (1991: 306) convincingly argue that "Household head data are biased socially toward the great generality of lives lived as household heads," but this statement is already loaded with the assumption that one person alone contributes the majority of income and energy to reproduce the household (see Katz, 2001 for a critique of this assumption). Privileging the individual characteristics of one specific member erodes the possibility of understanding the living arrangements and domestic relationships that produce and reproduce the household as a whole (Buzar et al, 2005, Yates, 2002; Wright and Ellis, 2006). Moreover, the gender breakdown of responses given by couples to Statistics Canada's requirement to identify a 'Person 1' in the Census questionnaire may be more reflective of the continuing purchase of the male breadwinner tradition in our collective imaginary than of the actual financial arrangements of attached individuals within a contemporary household. It is therefore not surprising that feminist researchers have been among the most attentive to the effects of demographic transformations in household composition, due in part to their preoccupation with themes such as the experience of everyday life, the idea of home, and gender (Buzar et al, 2005). With similar concerns in mind, I examine gender differences in household capacities of entry to homeownership by taking stock of the sex distribution of both primary household maintainers and, in the case of couples, of their spouses or common-law partners as well. (From here on 'spouses' will refer also to 91 common-law partners unless otherwise noted.) In order to do this, I shift the unit of analysis for this subsection of my study only, focusing on the individual instead of the household. In that respect, the figures I report here are inconsistent with the totals reported in the rest of this research, because they do not pertain to households but, rather, to individuals who are either Person 1 or a spouse. My goal is not to undermine the merits of exploring homeownership patterns of female-led households compared to male-led households (Edmonston, 2004), but rather to suggest a different avenue of investigation. Are women better or worse off than men in terms of their household's financial ability to obtain a mortgage, regardless of their status as, or relationship to, a primary household maintainer? Table 17 provides a summary of my findings for the sex variable. The first observation is that women who were heads of households or spouses in Montreal, Toronto and Vancouver were more likely than men to be tenants, as households with a male Primary Maintainer typically have higher rates of homeownership (Laryea, 1999). The larger numbers of women in all three C M A s suggests that they were also more likely to be heading households as unattached individuals than men, regardless of their C M A of residence (see for example Ley and Smith, 2000, Bunting et al., 2004). In contrast, Montreal had a much higher proportion of both women and men who would qualify for the average mortgage. One can clearly see here the effect of Montreal's lower house prices, and of average incomes that were catching up to Vancouver's (see Figures 1 and 2). But the aggregate income data used to produce those two introductory graphs masked the geographically uneven impact of gendered income inequalities on the capacity of tenant households to enter homeownership. Of all people seemingly shut out of the 92 Table 17: Mortgage-qualifying tenants, by sex of Person 1 and of spouse (if applicable) (Number of individuals and proportions of all tenants. Column percentage, 2001) Individuals Households - Female Person 1 or spouse (total # of individuals) Male Person 1 or spouse (total # of individuals) Total # of households Montreal: Total homeowners 643,800 585,000 708,450 Total tenants 526,700 418,500 709,850 Qualifying tenants, Group 1 ($45,000 and over) 27.6 35.6 26.6 Qualifying tenants, Group 2 ($35,000 to $44,999) 12.9 14.2 12.5 Non-qualifying tenants (Group 3, income < $35,000) 59.5 50.2 60.9 Toronto: Total homeowners 950,000 852,200 1,037,900 Total tenants 467,300 382,500 606,300 Qualifying tenants, Group 1 ($75,000 and over) 16.5 20.6 15.5 Qualifying tenants, Group 2 ($60,000 to $74,999) 10.1 11.6 9.6 Non-qualifying tenants (Group 3, income < $60,000) 73.4 67.8 74.9 Vancouver: Total homeowners 414,650 366,450 462,650 Total tenants 212,000 194,900 295,900 Qualifying tenants, Group 1 ($85,000 and over) 10.8 12.6 9.7 Qualifying tenants, Group 2 ($70,000 to $84,999) 7.2 7.5 6.3 Non-qualifying tenants (Group 3, income < $70,000) 82.1 79.9 84.0 Income figures are before tax for the year 2000, and apply to the household as a whole. For example, 59.5 percent of female primary household maintainers or female spouses of primary household maintainers in Montreal lived in households with total incomes of up to $34,999. Totals may not add up due to rounding. Source: Census 2001, P U M F (Individuals file). housing/mortgage market, women were the worse off: female primary maintainers or spouses were more likely than their male counterparts to be in a non-qualifying tenant household in 2001 (Table 17). This gender difference was particularly pronounced in Montreal and to a lesser extent in Toronto. In Vancouver, female Person 1 or spouses were only 2.2 percentage points more likely than their male counterparts to fall below the qualifying threshold. In Montreal, the gender rate differential was 9.3 percentage points, and 5.6 in Toronto. 93 Visible minority grouping The existence of the visible minority variable in the Census of Canada is tied to the decision by the federal government to monitor the incidence of discrimination by colour of skin in the hiring of public employees. In the housing market, evidence of this type of discrimination is mixed and has generally been critiqued for its methodological limitations (see Murdie et al., Forthcoming). This does not mean that discrimination based on colour of skin is non-existent in the housing sector, but simply that its incidence is very difficult to measure. The variegated geographies of immigrant settlement in metropolitan Canada have led to a different distribution of visible minority groups in the country's largest CMAs, but also to inter-metropolitan differences in housing and income outcomes among households headed by individuals belonging to same visible minority groups. In Toronto and Vancouver, for example, households in which Person-1 individuals identified as White had the second highest rates of homeownership after Chinese (Table 18). In the three cities, tenant households headed by Whites had some capacity of entry advantage over other tenant households, as they had the highest likelihood of qualifying for the moderate or average mortgage (at approximately the same rate as the C M A as a whole). But in Montreal and Toronto, Black PHMs had the highest likelihood of living in tenant households that did not qualify for a mortgage on moderate and average priced dwellings. In Vancouver, this ranking belonged to 'Other visible minorities' and, more surprisingly, to Chinese-headed tenant households. As Figure 19 in the Appendix indicates, Chinese-headed households had the highest likelihood of owning their own home in that C M A . This is another striking example of the inconsistency that can exist between the 94 Table 18: Qualifying tenant households, by visible minority category for Person 1 (Number of households and proportions of all tenants. Column percentages, 2001) Not a visible Other visible minority Chinese South Asian Black minority (White) Total Montreal: Total homeowners 8,750 5,150 11,250 18,350 665,000 708,450 Total tenants 9,000 12,350 36,550 47,950 603,950 709,850 Group 1 23.8 19.2 17.4 21.9 27.7 26.6 Group 2 9.0 15.9 10.7 12.5 12.6 12.5 Group 3 67.4 65.0 71.9 65.6 59.7 60.9 Toronto: Total homeowners 89,350 69,400 36,000 69,450 773,650 1,037,900 Total tenants 32,400 54,800 64,550 80,200 374,150 606,050 Group 1 11.8 11.1 8.0 14.2 18.0 15.5 Group 2 6.3 10.5 6.9 8.8 10.4 9.6 Group 3 81.9 78.4 85.1 77.0 71.6 74.9 Vancouver: Total homeowners 77,450 27,950 2,500 25,800 329,000 462,650 Total tenants 26,250 13,550 4,600 33,950 217,350 295,700 Group 1 6.2 7.6 7.2 5.4 10.9 9.7 Group 2 4.1 6.8 7.2 4.9 6.8 6.3 Group 3 89.7 85.6 86.0 89.6 82.3 84.0 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). homeownership attainment of one group and the capacity of entry into owner occupancy of its tenant households. (However, Hiebert et al. (Forthcoming) have found that the 2001 Census data for thousands of Chinese visible minority households contain incongruously low incomes relative to shelter costs. Accordingly, it may be that in this particular case, the above-mentioned inconsistency is at least partially the result of poor data.) It is also important to consider the fact that Chinese residents have long histories of settlement in Toronto and especially in Vancouver, but new waves of immigrants with changing characteristics have been arriving in large numbers since the 1970s. The category of visible minority is too broad to capture the variegated generational, place of 95 origin, and cohort characteristics that may be contributing to the disparity between high homeownership attainment for Chinese as a whole and low financial capacity to transition into owner occupancy for Chinese tenants in particular (Lo and Wang, 1997). Ethnicity (immigrants only) As in the case of visible minorities, the geographies of immigrant settlement in Canada yield different distributions of immigrant groups of diverse ethnic origins from one C M A to another. Groups that have high rates of financial accessibility to homeownership in one metropolitan area may not display the same level of economic achievement and purchasing power in others. For example, Vancouver tenant households in which Person-1 individuals were immigrants reporting East or South East Asian ethnic backgrounds were less likely to have sufficient incomes to qualify for a moderate mortgage compared to their East or South East Asian immigrant counterparts in Toronto and Montreal (Table 19). Multiple ethnicity immigrant PHMs fared worse in Montreal than in Toronto and Vancouver. In the three C M A s under analysis, the tenant households with the highest likelihood of not qualifying for the moderate and average mortgage had a P H M of African or West Asian ethnic background, followed by PHMs of Arab background in the case of Vancouver. These ethnic groups have among the smallest populations in their respective CMAs, so it is possible that the lack of institutional completeness and much smaller possibilities for intra-group social networking are negatively affecting tenant households. However, more favourable outcomes in the case of other small groups seem o rule out any straightforward relationship between population group size in a given C M A and household capacity to transition into homeownership. In fact, Lapointe and 96 Table 19: Qualifying tenant households, by ethnic category for Person 1, immigrants only (Number of households and proportions of all tenants. Row percentages, 2001) Montreal: Toronto: Vancouver: Owner hhlds. Tenant hhls. Group 1 Group 2 Group 3 Owner hhlds. Tenant hhls. Group 1 Group 2 Group 3 Owner hhlds. Tenant hhls. Group 1 " Group 2 Group 3 Total 1 4 0 , 0 5 0 1 5 6 , 0 5 0 2 2 . 4 1 2 . 3 6 5 . 2 5 4 8 , 8 5 0 3 2 2 , 0 5 0 1 2 . 5 9 . 0 7 8 . 4 1 9 9 , 5 0 0 1 0 6 , 0 0 0 7 . 5 5 . 5 8 7 . 0 European 8 1 , 6 0 0 4 9 , 2 5 0 2 7 . 7 1 0 . 6 6 1 . 8 2 5 4 , 3 5 0 9 0 , 5 0 0 1 4 . 0 9 . 6 7 6 . 4 6 1 , 2 0 0 2 8 , 1 5 0 8 . 0 5 .0 8 7 . 0 African 1 , 2 0 0 8 , 2 5 0 1 8 . 4 1 0 . 3 7 1 . 3 5 , 9 5 0 1 8 , 1 0 0 6 . 7 7 . 3 8 5 . 9 8 0 0 E 1 6 5 0 E 4 . 5 2 . 3 9 3 . 2 Arab 8 , 1 5 0 1 9 , 4 0 0 2 3 . 4 1 3 . 5 6 3 . 1 6 , 4 5 0 7 , 8 5 0 15.1 4 . 7 8 0 . 2 8 0 0 E 9 5 0 E 1 1 . 5 3 . 8 8 4 . 6 West Asian 4 , 0 0 0 5 , 7 5 0 17.1 1 1 . 6 7 1 . 3 9 , 4 5 0 1 3 , 2 5 0 8.1 5 . 9 8 6 . 0 2 , 4 5 0 5 , 1 5 0 4 . 3 2 . 9 9 2 . 8 South Asian 4 , 1 0 0 9 , 7 5 0 2 1 . 9 1 5 . 6 6 2 . 6 6 1 , 3 0 0 4 6 , 2 0 0 1 1 . 8 1 1 . 0 7 7 . 3 2 3 , 1 5 0 9 , 9 5 0 6 . 3 6 .7 8 7 . 1 East and SE Asian 1 4 , 0 0 0 1 6 , 1 0 0 2 5 . 1 1 0 . 6 6 4 . 3 1 0 6 , 8 5 0 5 6 , 9 0 0 14.1 7 . 9 7 8 . 1 8 0 , 2 5 0 3 6 , 3 0 0 5 .7 5 . 0 8 9 . 3 Latin, C , S. American 2 , 7 0 0 1 0 , 0 0 0 2 1 . 2 1 4 . 8 6 4 . 0 5 , 6 0 0 7 , 6 5 0 1 7 . 4 11 .1 7 1 . 5 7 0 0 E 2 9 5 0 E 8 . 8 7 . 5 8 3 . 8 Caribbean 7 , 8 5 0 1 7 , 3 0 0 1 7 . 5 1 2 . 2 7 0 . 3 2 7 , 3 5 0 3 4 , 5 0 0 8 . 5 7 .0 8 4 . 6 7 5 0 E . 7 5 0 E 1 0 . 0 1 0 . 0 8 0 . 0 Multiple ethnicities 1 6 , 5 0 0 2 0 , 2 0 0 1 4 . 6 1 5 . 4 70 .1 7 1 , 5 5 0 4 7 , 1 0 0 1 3 . 7 1 1 . 0 7 5 . 4 2 9 , 4 5 0 2 0 , 1 5 0 1 1 . 0 7 .2 8 1 . 8 F: Too unreliable to be published. E : Sample size for this category is small. Use this figure with caution. Totals may differ from other tables in this study due to non-response and rounding. Household incomes are before tax, for the year 2 0 0 0 . Source: Census 2 0 0 1 , P U M F (Individuals file). 97 Murdie's (1996) study of 1991 Census data suggests that household size and structure may be more closely related to these outcomes than ethnic group size, as high proportions of non-family single persons were found among households headed by immigrants from the Caribbean and Africa. Central, South and Latin American PHMs across the three C M A s live in households with slightly higher probabilities of qualifying for the moderate or. average mortgage than the C M A average, but they also have some of lowest rates of homeownership. Why are these households not converting their income potential into owner occupancy at faster rates, despite their higher tendency to be in households formed by married couples with children (Lapointe and Murdie, 1996)? Are remittances an important factor affecting homeownership accessibility for this group? Here again we encounter another area of homeownership accessibility research deserving more attention than it has received until now. A separate and very summary examination of the ethnic background variable for non-immigrants (not shown here) revealed that for Person-1 individuals in many ethnic categories, the likelihood of living in a tenant household that did not qualify for a moderate mortgage was slightly higher in the case of non-immigrants, particularly in Montreal and Vancouver. We know of course that some (not all) immigrant groups achieve higher rates of homeownership than the native-born after 15 or 20 years in Canada (Haan, 2005a; Lapointe and Murdie, 1996), but to my knowledge, the differential outcomes of ethnic groups by immigrant status have not been systematically examined in the Canadian housing literature. 98 Recent research has found that homeowner households in some visible minority groups have a higher likelihood of being at risk of loosing their homes due to low income than others, suggesting that the high level of sacrifice that average members of certain groups are prepared to make in order to achieve homeownership is indicative of a strong preference for owner occupation; for example, South Asian immigrant households in Vancouver—which are generally larger than the average immigrant household—appear to employ a strategy of income pooling within the household in order to achieve owner occupancy, while households in other visible minority groups seem less likely to exhibit such a pattern of housing consumption behaviour (Hiebert et al., Forthcoming). Other researchers have explicitly argued that ethno-cultural background plays a significant role in the decision-making process of some immigrant groups when it comes to home purchases (Haan, 2005b; Borjas 1998). While the methodology I employ in the present study is unable to either prove or disprove the incidence of an ethno-cultural effect in the rate of homeownership attainment, my findings do suggest that capacity of entry into owner occupancy is an additional factor worth considering when testing the validity of , such hypotheses. At the descriptive level of analysis, an examination of immigrant home-owning households, disaggregated by the visible minority or ethnic category (as well as age group and period of arrival) reported by Person 1, is well beyond the scope of the present study, but clearly constitutes an important avenue for future research. Immigration cohort Time of arrival can impact the homeownership rates of immigrants in at least two different ways: the first is that the longer immigrants live in Canada, the more likely they are to become homeowners (Haan, 2005a; Lapointe and Murdie, 1996; Laryea, 1999). 99 The second is that immigrants that arrive during an economic downturn may experience more difficulties during the initial adjustment period, which may in turn slow down the process of assimilation into the labour market (Borjas, 1985) and thus delay entry into owner occupancy (Haan, 2005a; Laryea, 1999). Table 20 examines the capacities of entry into homeownership of different immigration cohorts. In the three subject CMAs, tenant households in both the oldest and the most recent cohorts have the highest likelihood of not qualifying for a mortgage on the moderate and average priced dwelling posited in this thesis. While the outcome of the latter cohort is no doubt tied to recency of arrival, in the case of the former it is potentially far more worrisome, particularly in Vancouver where, once again, polarisation haunts the scene. Immigrants arriving before 1961 fall into two distinct sub-categories: those who immigrated as adults (and had for the most part already entered their retirement years by 2001), and those who arrived with their parents as children—the so-called generation one-and-half—and were therefore 40 years or older in 2001. A separate set of cross-tabulations for this cohort (figures not shown) revealed that in the three C M A s under study, it was the youngest PHMs in the cohort (the generation one-and-half ers between the ages of 40 and 54) who were by far the most likely to be in tenant households with non-qualifying income levels—considerably below the C M A average and recording negative spreads of between 15 and 20 percentage points with their parents and older siblings. This suggests that, at least for this cohort, the youngest among the one-and-half generation were more likely, as tenants, to be having a hard time catching up to the homeownership capacities of their parents and older siblings. Further age and cohort analysis would be required to verify this hypothesis. One may wonder, for 100 Table 20: Qualifying tenant households, by period of immigration of Person 1 ((Number of households and proportions of all tenants. Column percentages, 2001) Year of immigration: Arrived before 1961 1961-1970 1971-1980 1981-1990 1991-1995 1996-2001 All immig. cohorts Canadian-born Montreal: Total homeowners 39,400 36,250 32,550 19,900 9,600 4,100 141,850 565,350 Total tenants 15,900 15,750 25,400 37,200 28,800 35,050 158,100 538,800 Group 1 21.3 30.9 27.4 • 22.7 24.3 18.3 23.4 27.7 Group 2 8.3 11.5 10.9 15.3 13.4 10.9 12.2 12.8 Group 3 70.4 57.6 61.6 62.0 62.4 70.8 64.4 59.5 Toronto: Total homeowners 117,400 113,150 122,050 101,350 61,400 37,350 552,750 481,850 Total tenants 24,900 29,800 50,600 68,350 '66,500 81,950 322,050 270,400 Group 1 10.4 16.2 16.3 12.8 12.5 9.2 12.5 19.2 Group 2 7.3 10.2 9.1 9.6 9.2 8.5 9.0 10.5 Group 3 82.4 73.6 74.5 77.6 78.3 82.2 78.4 70.4 Vancouver: Total homeowners 37,550 31,700 44,200 35,500 32,900 19,800 201,700 259,600 Total tenants 10,550 10,650 16,850 20,550 18,300 29,100 106,000 181,050 Group 1 5.2 8.3 12.1 7.4 8.9 4.4 7.5 11.1 Group 2 2.8 8.6 7.5 6.1 . 5.5 3.8 5.5 7.0 Group 3 91.9 83.1 80.5 86.5 85.7 91.8 87.0 82.0 Immigrant household figures do not include non-permanent resident households. Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). instance, why was the 1961-1070 cohort (another double-generation one) so much better off than most others in 2001? Is it because the first-generation PHMs had reached the age of highest ownership rates between 1996 and 2001, right as their children were entering their prime first-home buying age at a time of relative low home prices and low interest rates? Generational ties to immigration Unlike immigrants whose children came with them from abroad, those with children born in Canada tend to have "an older age structure than the non-immigrant 101 population because their Canadian-born offspring become non-immigrants." (Lapointe and Murdie, 1996: 15). An older age structure in turn tends to translate into higher ownership rates for the group. We know that, other things held constant, households generally experience an increase in the probability of living in owner occupied housing as their P H M grows older but remains below retirement age. Yet first-generation immigrants have the highest proportion pf tenant households that would not have qualified for this study's target mortgages (Table 21). This is no doubt due to the large proportion of recent immigrants making up the first generation. But the significance of this finding is the absolute (rather than relative) size of this population, especially in Toronto, where there were more than 265,000 tenant households in Group 3. A certainly more insightful finding is that having a Canadian-born parent gives second-generation tenant households a slightly better likelihood of meeting the qualifying income threshold in Toronto and Vancouver, although the opposite is the case in Montreal. This last result suggests that it is unlikely that, everything being equal, the place of birth of a Person l 's parents bears any significant effect on the capacity of entry into homeownership of second-generation tenant households. But this does not discountthe possibility that generational ties to immigration may play an interaction effect with other socio-demographic or place characteristics, helping to explain the discrepancy of results between Montreal on the one hand and Toronto and Vancouver on the other. The experiences and general outcomes of individuals and households with some generational detachment from immigration require more detailed investigation i f we are to understand the potential inter-generational impacts of the inability to access 102 Table 21: Qualifying tenant households, by generational ties to immigration of Person 1 (Number of households and proportions of all tenants. Column percentages, 2001) 2nd generation: 1st generation both parents born outside Canada 2nd generation: one parent born outside Canada 3rd generation and over Total Montreal: T o t a l h o m e o w n e r s 143,600 32,500 • 28,850 503,550 708,450 T o t a l t e n a n t s 172,450 31,150 28,650 477,600 709,850 G r o u p 1 23.2 32.2 27.1 27.4 26.6 G r o u p 2 11.7 11.9 13.6 12.8 12.5 G r o u p 3 65.0 56.0 59.3 59.8 60.9 Toronto: T o t a l h o m e o w n e r s 558,250 125,500 74,350 279,850 1,037,900 T o t a l t e n a n t s 337,300 52,550 39,750 176,500 606,050 G r o u p 1 12.6 16.5 18.8 20.0 15.5 G r o u p 2 8.8 9.1 10.8 10.9 9.6 G r o u p 3 78.6 74.3 70.4 69.1 74.9 Vancouver: T o t a l h o m e o w n e r s 204,700 59,750 49,850 148,350 462,650 T o t a l t e n a n t s 115,400 30,400 30,750 119,150 295,700 G r o u p 1 7.4 10.2 12.6 10.9 9.7 G r o u p 2 5.4 6.1 6.3 7.3 6.3 . G r o u p 3 87.2 83.6 81.1 81.8 84.0 Totals may not add up due to rounding. Household incomes are before tax, for the year 2000. Source: Census 2001, P U M F (Individuals file). homeownership at all or else at an early enough stage in the life cycle. Taken together, these cross-tabulations suggest that insufficient income is the main reason why the majority of tenant households are not homeowners. My analysis of key 2001 Census variables, broken down into detailed socio-demographic subcategories, reveals that the proportion of all tenant households with most difficulty gaining entry into homeownership was almost always greater than 50 percent and very often above 75 percent, particularly in the expensive housing markets of Toronto and Vancouver. With respect to tenant households, it would appear that economic class trumps other demographic factors when it comes to explaining the (in)ability to enter homeownership. As we have seen, this is especially the case when polarisation between homeowners and tenants is apparent within particular subgroups. This last finding requires some qualifications. First, the results upon which this insight is based rest on the assumption that the socio-demographic characteristics I have analysed here are the most pertinent, and that the data describing such characteristics are consistent and reliable. Secondly, all my results are coloured by the definitions that I adopted from the outset, particularly with respect to target house prices. Based on informed opinion, I believe these definitions are reasonable, but this is ultimately a subjective judgement, open to independent verification. Finally, due to the descriptive nature of the bi-variate (and some cases 'tri-variate') analysis I conducted, interaction effects were not taken thoroughly into account. Clearly, these may have a significantly debilitating impact on the explanatory power of the most interesting variables in this study. Despite these caveats, it is worth noting that many of the categories included in this study revealed what appears to be a high level of polarisation within certain sub-groups. That is, several subpopulations not only displayed high rates of homeownership, but also recorded extremely small proportions of tenant households in Groups 1 and 2. As Figures 10, 15, 16, 19, 22, and 23 in the Appendix show, this situation was particularly noticeable in Toronto and even more so in Vancouver. There are at least two ways of interpreting these results. On the one hand, it is important to remember that many home-owning households reported income levels that were comparable to those of Group 3 tenants (Table 7). From this perspective, homeownership appears to simply hide a large 104 number of low-income households, suggesting that housing tenure is not the threshold of polarisation that it appears to be. On the other hand, it is clear that what sets home-owning households apart is the equity that tends to accumulate on a residential property, particularly in times of rising dwelling prices. A Toronto household that purchased an average priced dwelling in 2001, for example, would have seen the property's real price grow by $46,000 by 2004, assuming it retained its average-price standing in the market (Table 22). Even though this figure represents the lowest average-price appreciation in real dollars among the three CMAs, it is clearly higher than the cost of carrying a conventional mortgage or the interest that would have been accrued on an amount of money equivalent to a 10-percent down payment during that period. Three years after the home purchase (other things being equal), such a household would have been thousands of dollars on top of other households that had similar incomes but lacked the asset accumulation potential associated with homeownership. I do not wish to deny the serious risks faced by low-income home-owning households that spend large portions of their income trying to retain possession of a Table 22: Average dwelling price appreciation, in 2004 constant dollars, 2001 to 2004 2001 2002 2003 2004 Average dwelling price Average dwelling price Price difference, 2001 to 2002 Average dwelling price Price difference, 2001 to 2003 Average dwelling price Price difference, 2001 to 2004 Montreal 134,600 150,350 15,750 170,200 35,600 194,700 60,100 Toronto 269,250 288,850 19,600 298,800 29,550 315,250 46,000 Vancouver 306,050 315,650 9,600 335,650 29,600 373,900 67,850 Note: Average yearly sales prices as per the Canadian Real Estate Association's Multiple Listings Services (MLS) for each Metropolitan area, reported by the C M H C and adjusted by the author using Canadian average Consumer Price Index figures. CREA's definition of Metropolitan Area differs from Statistics Canada's definition of CMAs . Sources: CMHC/Canadian Real Estate Association; Statistics Canada. 105 purchased home. As discussed in Chapter 2, not everyone benefits equally from owner occupancy, especially when the location of the purchased property or the time at which the acquisition was made are not optimal. Nonetheless, the magnitude of the Group 3 subgroups revealed in this study suggests that, generally speaking, homeownership attainment is a sharp marker of wealth (if not income) differences between owner-occupying households and tenant households lacking major assets, particularly for certain sub-groups in expensive housing markets such as Toronto and Vancouver. The potential long-term consequences of this tenure-driven socio-economic gap should give pause to anyone who sees nothing but good news in the long, ongoing rally of escalating house prices. 106 DISCUSSION AND C O N C L U D I N G R E M A R K S I began this thesis with a description of how a housing boom took over the real estate markets of Canada's largest metropolitan centres at the turn of the new millennium, raising dwelling prices at faster rates than household incomes. I proceeded to identify the individual and social benefits of rising personal-property prices, as well as many of the corresponding negative effects, emphasising in particular the growing gap in income and wealth levels between homeowners and tenants. I followed this with a discussion of the difficulties associated with attributing causality around such problems and with defining and measuring affordability. I then commented on the advantages and limitations of the two most important housing affordability measures used in the academic and policy literature in Canada, the Core Housing Need ratio and the In Need/At Least Half measure. This critical discussion was summed up by an explanation of my choice of the terms 'financial accessibility' and 'capacity of entry' as more specific expressions of the type of affordability problem that concerns me here. The second chapter of this thesis reviews the reasons why it is important not to neglect the study of opportunities and barriers to upward mobility of housing tenure, a current research lapse that may be the unintended result of prioritising household risk measures associated with the loss of current housing or tenure status, which seems to be the case with the Core Housing Need and the INALH measures favoured by the Canadian federal government. The inability of a sizeable minority of the population to live in owner occupied housing is again placed in the context of the enhancement of polarisation that results from the relationship between owner occupancy and residential tenancy as it has evolved over the past half-century in Canada. The rest of the chapter provides an 107 overview of my methodology and the multiple assumptions that were required to conduct the empirical analysis at the heart of this study, followed by a brief overview of the metropolitan geographies of my three research sites, namely Montreal, Toronto and Vancouver. The third and final chapter of this thesis begins by providing an empirical evaluation of the considerable difference between C M H C ' s affordability measure and the accessibility index I advocate in this study. It then moves to the more extensive presentation of tenant household profiles emerging from my cross-tabulations of 2001 Census data. These profiles are expressed in the form of proportions of tenant households with given socio-demographic characteristics and which, everything else held constant: a) would have been financially able to qualify for a conventional mortgage on an average priced home in their C M A of residence (labelled Group 1); b) would have been financially unable to qualify for a conventional mortgage on an average priced dwelling in their C M A of residence but would have qualified i f the dwelling was moderately priced (labelled Group 2); and c) would have been financially unable to qualify for a conventional mortgage on either an average or moderately priced dwelling in their C M A of residence (labelled Group 3). This classification was of course conducted within the framework established by the conservative down payment conditions and equally conservative definitions of conventional mortgage and average and moderate dwelling prices that I adopted for the purposes of this study. 108 One of the most important findings of this thesis is the observation that the homeownership patterns of population sub-groups do not always travel well to the realm of capacities of entry into owner occupancy. For example, Chinese visible minorities were found to have the highest rates of homeownership in the three C M A s under study, but were also found to be the tenant group with the lowest probability of qualifying for the moderate or average mortgage. Besides suggesting a highly polarised group composition (if not in terms of income, certainly in terms of gross equity accrued), this result seems to indicate that research on homeownership attainment alone cannot be relied on for the purposes of studying the accessibility of owner occupancy as a matter of tenure transition capacity. Haan (2005a) recently found that Census variables that can explain homeownership attainment do not necessarily have the power to explain changes in homeownership levels among groups; he concludes that approximately two-thirds of the changes in the homeownership attainment rates of immigrants and non-immigrants between 1981 and 2001 cannot be accounted for by the Census variables expected to have an effect. My research suggests that at least part of the unexplained changes in homeownership attainment levels may be related not simply to income, geography and demographic factors in general, but also to the income distribution of tenant households along a CMA-specific hierarchy of financial accessibility to homeownership, which is itself tightly related to metropolitan incomes, house prices, and current conditions of borrowing. Simply put, a sub-group's rate of homeownership is not necessarily a good predictor of the ability of its tenant members to transition into owner occupancy. 109 My analysis of tenant profiles also led to several insights regarding the potential relationships between several socio-demographic characteristics of tenant households and their capacity of entry into owner occupancy. These insights are summarised below. Insights gained from analysis of the 2001 Census data In the first place, my results suggest that geography and economic class play an important role in determining the ability of many households to access owner occupancy. A third critical factor is the importance of cohort effects. Entering prime first-home buying age during times when home purchasing is more difficult (due for example to economic downturns or to dwelling price booms) seemed to delay-the ability of tenant households headed by individuals in that age group from catching up to the homeownership levels of other households. Secondly, my research found that various changes in societal trends could potential act as key determinants of the capacity of entry into homeownership. Recent shifts towards smaller household sizes appear to delay or make it more difficult for tenant households to be financially able to access homeownership due to smaller incomes compared to households with more wage earners. Similarly, the growing need for both parents to contribute towards the household's finances may play a role in shutting many households with young children out of the homeownership market. I speculated that the unavailability of affordable childcare may cause such households to divert financial resources that could be used towards achieving owner occupancy, at least until the children are old enough. 110 Another example of the barriers to homeownership emerging from recent transformations in the once-typical household composition relates to the continuing unevenness in the gender distribution of labour and wages. The lower wages women tend to perceive in the job market make them increasingly vulnerable to being excluded from the equity building potential of homeownership, especially in the case of single mothers and lone heads of household. Thirdly, the education achievement of the Primary Household Maintainer (PHM) was also found to affect the probability that a tenant household is financially capable of accessing homeownership. My study found that households with higher education were not always capable of entering homeownership despite their relatively high incomes. As a possible explanation, I proposed that large student loan debt-loads were absorbing large portions of income in many of these households, a situation that can last for many years after graduation. A l l of the above socio-demographic characteristics are associated with the life-cycle model; but for the purposes of this summary of findings, I need to introduce a second set of variables employed in my study, pertaining to the characteristics of immigrants. The first of these characteristics is household size, as households where the P H M are immigrants are on average larger than those in which Person 1 is native born. Consistent with other research, multiple family households generally have higher household incomes, suggesting that they have more members who contribute financially to running the household. As a result of their higher income, they were found to be more capable of accessing homeownership than smaller households with single-family compositions. In the case of South Asian households, a high rate of homeownership in i l l Vancouver provides some evidence in favour of this hypothesis Unfortunately, this strategy for homeownership attainment has the effect of diluting any potential accumulated equity among a larger number of individuals. But home purchasing is not the priority for all immigrants, and other incentives may propel newcomer tenant households to integrate as many of their members as possible into the labour market. One such incentive is the commitment to send money to relatives overseas, especially during the first years of settlement. While having more members contributing to the household's income should prove beneficial, sizeable transnational outlays in the form of remittances can delay entry into homeownership. Households led by Central or South American PHMs seem to fall into this situation, as their rate of homeownership was low but they had a relatively high proportion of tenant households in Group 1 and 2. In my analysis of visible minority or ethnic background data, I found several trends in the likelihood of homeownership attainment and accessibility of entry among specific subgroups, as in the case of the two examples mentioned above. But the limited nature of my research design renders tentative any conclusion regarding the role of ethno-cultural characteristics in determining the housing outcomes of immigrants. There are simply too many factors creating a potential interaction effect with these two variables, including those that relate to the opportunity structures typically associated with social networks; a simple bi-variate descriptive analysis is clearly inadequate to capture the complexity of this much-debated issue. Previous research has suggested that qualitative research is a necessary complement to statistically based analysis of such demographic 112 characteristics. The results presented in this thesis might assist in the design of such types of study. Another immigrant characteristic worth mentioning in this summary of findings is knowledge and home use of an official language. This was found to be an issue for immigrant households that were unable to attain homeownership rapidly following arrival. For these tenant households, the degree of acquisition of an official language was generally found to go hand in hand with the financial capacity of entry into homeownership. One can speculate that this outcome is tied to the broadening of employment options and the potential expansion of social networks. The final two insights extracted from this research pertain to the generational ties to immigration of the P H M . It would appear that the non-immigrant status of one of the parents of a second-generation householder positively affects the likelihood of homeownership accessibility of the second generation, although this result could be linked to unknown interaction effects. Similarly, an interaction effect may have contributed to diminished levels of homeownership accessibility for native-born PHMs whose parents are both immigrants. In turn, PHMs who are members of the 'one-and-half generation appear to be at a disadvantage over their parents and over native-born children of immigrants. One possible explanation is that these individuals do not necessarily receive wealth transfers from their parents as a 'hand-up' that makes access to homeownership possible, because first-generation immigrants generally begin accumulating net housing equity only after numerous years of settlement. 113 While the results presented in this study are of a descriptive nature and do not purport to establish causality, they do provide substantial material for further research on the consumption capacity and behaviour of households with regards to housing. Most of the insights that were briefly reported in this concluding chapter are related to household and P H M features that change or can change over the lifespan of the household (i.e., household composition, household size, presence of young children in the household, number of household members with paid employment, age of Person 1,'highest level of education of Person 1, and command of an official language by Person 1), while others are derived from household and P H M characteristics that normally do not change over time (i.e., sex of Person 1 or spouse, and generational relationship to immigration). This observation is meant to call attention to the fact that individuals, households and society at-large are likely to deal with immutable socio-demographic characteristics in a different way than they would with features that can and do change throughout a person's life cycle. Simply put, the interplay of these two sets of variables provides a framework in which households are able to exercise their agency and even surpass many structural constraints. Multivariate analysis can be used to estimate any amplifying or mitigating effect that the latter set of variables may have over the former and vice-versa. Qualitative research that employs questionnaire-based surveys can also help to develop a more complete understanding of how these diverse factors affect the opportunities and lived experience of 'priced-out' tenant households. Looking at the boom years The main goal of this study was to develop a profile of tenant households that were already priced out of owner occupancy in 2001— the year in which market 114 conditions were the most favourable for buyers in the 10-year period stretching from 1996 to 2005, but also the year before the price boom decidedly took off in metropolitan Canada. To facilitate analysis and discussion, I defined as 'priced out' those tenant households with insufficient incomes to qualify for a conventional mortgage on what I described as a moderately priced dwelling, and assigned the label 'Group 3' to them. At the most general level, my findings can be summarised with a C M A breakdown of the more than one million tenant households that fell within this category (Table 23). The figures presented in Table 23 are sobering in their magnitude, yet they must, be considered in the context of the limitations of my study. Bearing in mind the large number of home-owning households who, based on income alone, would not have qualified for a moderate mortgage in 2001 (Table 7), and in view of the strong sales volumes that propelled the price boom from 2002 onwards, it seems likely that many Group 3 tenant households have in fact been able to achieve homeownership in the years following the Census of 2001. There are several possible reasons why my estimates of the proportion of households with most difficulty accessing homeownership may not reflect the actual home buying capacity of households classified as Group 3. For the most part, these Table 23: Total number of non-qualifying tenant households, 2001 Montreal Toronto Vancouver Total number of tenant households belonging to Group 3 432,300 453,950 248,400 Sources: 2001 Census P U M F (Individuals file); C M H C (2005); 2001 Survey of Household Spending (Table 62F0035); Bank of Canada (CANSIM Table V122521). 115 explanations relate to the various assumptions underlying my analysis. For example, the house price figures that I employed in this study assume that a moderately priced dwelling was equivalent to approximately 75 percent of the average price, which was found to correspond to the minimum price recorded by approximately two thirds of all I MLS-listed sales in Toronto and Montreal in 2001. This means that one third of all M L S -listed sales were closed at prices below the moderate level, making homeownership accessible to an upper tier of Group 3 tenant households (assuming those households are interested in purchasing such units, and that they are able to find them). There is also the possibility that the income information contained in the Census data is flawed, due perhaps to either deliberate or unintentional under-reporting (see Hiebert, 2002; Hiebert et al, Forthcoming), which would obviously lead to an overestimation of the number of tenant households with most difficulty entering homeownership. Further research comparing results between various data sources, including the Survey of Consumer Finances, the Survey of Household Spending, and the Longitudinal Survey of Immigrants to Canada, could help determine the significance of such under-reporting should its existence be confirmed. Aside from issues of accuracy in the income data employed, one of the main challenges affecting research on the capacity of entry into owner occupancy is the lack of comprehensive and reasonably accessible data on household savings and wealth. (For a fee, Statistics Canada provides cross-tabulations of Financial Data and Charitable Donations compiled from annual tax returns submitted to the Canadian Revenue Agency, but the prohibitive cost of ordering these, and the inadequate degree of socio-demographic detail available for cross-tabulation constitute a serious limit to the I 116 usefulness of this data source for the type of research I am conducting.) Earnings derived from interest and investments could indeed be high enough to generate enough cash to make a substantial down payment, and also to lift a household above the mortgage-qualifying thresholds defined in this study. However, the number of tenants that derive large amounts of cash from their rents is unlikely to be very large. Therefore, the inclusion of earnings information (if it can be accessed) is expected to have but a minor impact on the results I have reported here. Yet, a final and perhaps a more important form of data deficiency in the context of the present research is the lack of information on the social capital of households. If tenant households do not have the financial capacity to enter homeownership themselves, they might be able to access the income and savings of others through their social networks. Unfortunately, the Census does not collect information on these types of support structures. (In the case of recent immigrants, one possible option for data on income, savings, detailed socio-demographic characteristics, and social networks is the Longitudinal Survey of Immigrants to Canada.) Based on an analysis of Statistics Canada's Survey of Household Spending from 1997 to 2003, Derrick Thomas (2005: 3.7) argues that buyers became more "atypical" in 2002, posting higher than usual rates of single, divorced or separated individuals (who tend to have lower incomes). He also found that higher percentages of buyers were indicating that their new home was in need of major repairs. In addition, some of the boom period buyers may be new households formed by individuals who lived with their homeowner parents instead of as a separate household. These PHMs may have benefited from an early inheritance, made possible by the appreciation of their parents' dwelling and by low interest rates that encourage re-financing and further borrowing against 117 accumulated home equity. It is worth noting too that consumer behaviour is changing, perhaps in response to an increasingly expensive housing market. Over the last decade, high prices, smaller household sizes and changing housing preferences have been combining to channel some households into ownership of small units (Lapointe & Murdie, 1996). Thomas (2005: 3.7) also found that first-time buyer households (those that fulfil the minimum income threshold to enter owner occupancy) did not display higher probabilities of home purchasing during the first years of the boom, because for these households "income level ... has more influence on the size and price of the house purchased than on the probability of buying a first house." Thomas (2005: 3.6) attributes the housing boom instead to low interest rates ("each percentage-point drop in interest rates between 1997 and 2003 spurred about 16,000 additional first-time home buyers") and their interaction with three main demographic factors: "the maturation of the baby-boom-echo, declining household size and immigration." (2005, 3.8) At least two additional factors are worth considering here. The first one is the home buying frenzy that occurred in 2002, as mortgage rates began to slowly rise from their 2001 low of 4.6 percent. While Thomas does not mention this phenomenon, he does provide some support for the argument that mass psychology played a role in fuelling the boom after 2001. In 2002, the real median household income of buyers according the Survey of Household Spending was $63,140, which is more than $4,000 below the median in the five years prior and almost $7,000 lower than it was in 2003 (2005: 3.7). Most households that purchase housing do so with debt (Carter, 2004), and the proliferation of low-income mortgage instruments (Hackworth and Wyly, 2003) has no 118 doubt had an important effect on the ability of many tenant households to qualify. A second factor not addressed by Thomas is the role that in-suite rental-income potential plays in facilitating access to mortgage capital for aspiring home-buyer households. This is particularly relevant in cities like Vancouver and Toronto, where house prices are the highest and a large share of the stock is made up of single detached units that can be retrofitted with a separate suite in the basement or attic. There is a final reason why my study does not seem to 'match the facts' in terms of housing sales volumes during the boom years. My research design clearly underestimated the importance of interaction effects, especially with regards to temporal variables like Person-1 age group and time of arrival (in the case of immigrant PHMs). In retrospect, the descriptive analysis that I conducted in this thesis would have benefited from the incorporation of the age group variable in all the cross-tabulations I performed. This approach would have allowed me to distinguish between tenant households that were below the peak home buying age and those that were at or above it. In other words, I would have been able to quantify more precisely those tenant households with a higher likelihood of being shut out of owner occupancy in the medium and long term, based on their life cycle stages and the responses they provided in the Census questionnaire. Assuming that the distribution of tenure and homeownership accessibility for the age group with the highest rate of homeownership (i.e. 55 to 64 years of age for Montreal and Toronto, and 65 to 79 years for Vancouver) is a good predictor of the proportion of households that are the least likely to attain homeownership over their life time, it is possible to estimate the total number of 2001 PHMs that are likely to suffer long term inaccessibility to homeownership (Table 24). A similar type of simulation could be 119 Table 24: Estimated number of households most likely to suffer long-term inaccessibility to homeownership, 2001 Total Rate of . potential long-term inaccessibility* 15-24 25-34 35-44 45-54 55-64 65-79 8 0 a n d over Number of households As % of all CMA households Montreal 24.4% 10,670 20,077 19,716 16,089 12,246 19,586 7,072 105,452 7.4% Toronto 19.0% 4,745 20,626 22,805 14,615 8,520 10,637 4,134 86,048 5.2% Vancouver 22.0% 4,451 14,676 13,480 9,681 4,626 5,456 2,355 54,711 7.2% * The rate of potential long-term inaccessibility to homeownership is equivalent to the percentage of Group 3 tenant households in the peak home-owning age group for each C M A , expressed as a proportion of all households in that age group. In this table, the rate has been applied to Group 3 households in each subgroup. Source: Household income data, 2001 Census P U M F (Individuals file). performed in the case of immigrant tenant households, using the variable 'period of arrival.' The calculations used to derive these figures (and the many others contained in ' this thesis) can no doubt be re-worked and refined so as to address many of their limitations and the problems deriving from formulated and underlying assumptions. While I recognise the mismatch between my descriptive, Census-based methodology and the actual strategies and processes of homeownership attainment that households adopt based on place of residence, household situation, and local market conditions, I believe that the results presented in this thesis remain useful in their present form, i f only because they constitute an early effort to measure and assess the intensity of the entry problem in Canada's metropolitan housing markets at the turn of the twenty-first century, and because of the insights it provides on its demographic variegation and its unevenness in terms of geographical location. Methodological debates about the proper measurement of affected tenant populations need to be fostered and welcomed, but they should not 120 override the need to understand the various types of obstacles to tenure mobility and how they affect a wide diversity of population sub-groups. A pessimistic reading of this study's findings would likely hold on to the fact that not all households benefit equally from homeownership, and argue that other types of housing programmes that do not privilege owner occupancy need to be set up in order to provide security of tenure to lower income tenant households. At the opposite extreme, the 'glass is half-full' reading would perhaps fasten on to the notion that currently priced-out households effectively constitute an untapped market segment of enormous proportions. Indeed, tenant households in 2001 would have represented a metropolitan home-buying capacity of $26.7 billion in Montreal, $28.7 billion in Toronto, and $10.1 billion in Vancouver. By metropolitan home-buying capacity, Listokin et al., 2002 mean the amount of sales, in dollars, that would be realised i f all tenant households had the financial capacity to purchase an adequate dwelling in their C M A of residence at a given price (which in this case corresponds to the moderate price levels defined in this paper). Neither of these two readings is completely off the mark, but in my opinion, a more balanced assessment of what this study reveals would not only need to consider the housing needs of the large minority of households that find themselves excluded from owner occupancy (and of the market opportunities that this provides), but also to question the contemporary version of the 'wil l to possess' and the polarising side-effects it can have, particularly at times of rising home prices (Table 22). The higher the exchange value associated with an owner-occupied dwelling, the larger the gap in gross wealth between that homeowner and the numerous tenant households that are priced out of the 121 market. In the absence of sufficient and reliable economic safeguards for all metropolitan residents, this formula becomes a matter of grave social concern. The relationship between shelter needs and homeownership has become increasingly complex in the past few decades, encouraging an increasingly conflicted view of housing—one in which dwelling is no longer conceived only as home, but more and more as commodity and investment. In a context where economic security is not universally and adequately guaranteed, this ideological shift is effectively promoting a market guilty of enhancing polarisation over time. I hope this thesis has contributed to an understanding of the socially and geographically uneven outcomes and reactions that such a relationship foments, and why this is something that urgently requires a collective effort to address. 122 WORKS CITED Alba, R. and J. Logan (1992) "Assimilation and Stratification in the Homeownership Patterns of Racial and Ethnic Groups." International Migration Review, 26: 1314-1341. Bacher, J. (1993) Keeping to the Marketplace. The Evolution of Canadian Housing Policy. Montreal: McGill-Queen's University Press. Barnes, T., D. Edgington, K. Denike, and T. McGee (1992) "Vancouver, the Province, and the Pacific Rim." In G. Wynn and T. Oke (eds.) Vancouver and its Region. Vancouver: U B C Press, 171-199. Boehm, T. and A . Schlottmann (2002) "Housing and Wealth Accumulation: Intergenerational Impacts." In Retsinas, N . and E. Belsky (eds.) Low-Income Homeownership. Examining the Unexamined Goal, 407-426. Cambridge, Mass. & Washington, D.C.: Joint Center for Housing Studies & The Brookings Institution. Boehm, T. and A. Schlottmann (1999) "Does Home Ownership by Parents Have an Economic Impact on Their Children?" Journal of Housing Economics, 8(3): 217-32. Borjas, G. (1998) "To Ghetto or Not to Ghetto: Ethnicity and Residential Segregation." Journal of Urban Economics, 44, 2: 228-253. Borjas, G. (1985), "Assimilation, Changes in Cohort Quality, and the Earnings of Immigrants." Journal of Labour Economics, 3: 463 - 489. Bourne, L. (1981) The Geography of Housing. New York: John Wiley & Sons. Bunting, T., A . Walks, and P. Filion (2004) "The Uneven Geography or Housing Affordability Stress in Canadian Metropolitan Areas." Housing Studies, 19(3): 361-393. Burke, T. and D. Hayward (2001) "Melbourne's Housing Pasts, Housing Futures." Urban Policy and Research, 19(3): 291-310. Buzar, S., P. Ogden, and R. Hall (2005) "Households Matter: The Quiet Demography of Urban Transformation." Progress in Human Geography, 29(4): 413-436. Carter, T. (2004) "Canadian Housing Prices: The Road to Wealth?" Urban Policy and Research, 22(1): 35-48. Carter, T. (1997) "Current Practices for Procuring Affordable Housing: The Canadian Context." Housing Policy Debate, 8(3): 593-631. Chaplin, R. and A . Freeman (1999) "Towards an Accurate Description of Affordability." Urban Studies, 36(11): 1949-1957. Chinloy, P. (1999) "Housing, Illiquidity, and Wealth." Journal of Real Estate Finance and Economics, 19(1): 69-83. Chui, T. (2003) Longitudinal Survey of Immigrants to Canada: Process, Progress and Prospects. Ottawa: Statistics Canada. 123 C M H C (2005) Canadian Housing Observer. Ottawa: Canada Mortgage and Housing Corporation. C M H C (2004) Research Highlight. 2001 Census Housing Series Issue 7 Revised: Immigrant Households. Ottawa: Canada Mortgage and Housing Corporation. Demographia International (2006) 2nd Annual Demographia International Housing Survey: 2006. Rating for All Major Urban Markets. Last accessed June 29, 2006: http://www.demographia.com/dhi-ix2005q3.pdf DeVoretz, D. (2004) Immigration Policy: Methods of Economic Assessment. Working Paper 04-13. Vancouver: Vancouver Centre of Excellence, Research on Immigration and Integration in the Metropolis. Doucet, M . and J. Weaver (1991) Housing the North American City. Montreal & Kingston: McGill-Queen's University Press. Economist (2005) "Hear that hissing sound?" The Economist, December 5: 86. Economist (2002) "Going Through the Roof." The Economist, March 28: 56-61. Edmonston, B. (2004) "Who Owns? Homeownership Trends for Immigrants in Canada." Paper presented at the Annual Meeting of the Canadian Population Society. Winnipeg, June 1-5. Engeland, J., R. Lewis, S. Ehrlich, and J. Che (2005) Evolving Housing Conditions in Canada's Metropolitan Areas, 1991-2001. Ottawa: Statistics Canada. Foote, N . , J. Abu-Lughod, M . Foley and L. Winnick (1960) Housing Choices and Constraints. New York: McGraw-Hill. Ford, J., R. Burrows, and S. Nettleton (2001) Home Ownership in a Risk Society. A Social Analysis of Mortgage Arrears and Possessions. Bristol: The Policy Press. Forrest, R., A . Murie, and P. Williams (1990) Home Ownership. Differentiation and Fragmentation. London: Unwin Hyman. Freeman, A. , C. Kiddle and C. Whitehead (2000) "Defining Affordability." In Monk, S. and C. Whitehead (eds) Restructuring Housing Systems: From Social to Affordable Housing?, 42-49. York: York Publishing Services. Gabriel, M. , .K. Jacobs, K. Arthurson, T. Burke, and J. Yates (2005) Conceptualising and Measuring the Housing Affordability Problem. Melbourne: Australian Housing and Urban Research Institute. Last accessed: July 7, 2006: http://www.ahuri.edu.au/downloads/NRV3/NRV3_A_Conceptualising_Housing_ Affordability.pdf Germain, A. (1999) "Les quartiers multiethniques Montrealais: une lecture urbaine." Recherches sociographiques, 40: 9-32. G V R D (2006) Regional Affordable Housing Strategy Workshop: Issues and Options. Affordable Housing Issues Paper (February 24, 2006), Greater Vancouver Regional District. Last accessed May 2, 2006: http://www.gvrd.bc.ca/growth/ workinggroup/LRSPReview-AffHsgWrkbk-final.pdf 124 Gyourko, J., P. Linneman, and S. Wachter (1999) "Analyzing the Relationships Among Race, Wealth, and Home Ownership in America." Journal of Housing Economics, 8(2): 63-89. Haan, M . (2005a) The Decline of the Immigrant Homeownership Advantage: Life-Cycle, Declining Fortunes and Changing Housing Careers in Montreal, Toronto and Vancouver, 1981-2001. Ottawa: Statistics Canada. Haan, M . (2005b) Are Immigrants Buying To Get In? The Role of Ethnic Clustering on the Homeownership Propensities of 12 Toronto Immigrant Groups, 1996-2001. Ottawa: Statistics Canada. Hacking, I. (1983) Representing and Intervening: Introductory Topics in the Philosophy of Natural Science. Cambridge, England: Cambridge University Press. Hackworth, J. and E. Wyly (2003) "Social Polarization and the Politics of Low Income Mortgage Lending in the United States." Geografiska Annaler, 85B(3): 149-165. Heisz, A . (2005) Ten Things to Know About Canadian Metropolitan Areas: A Synthesis of Statistics Canada's Trends and Conditions in Census Metropolitan Areas Series. Ottawa: Statistics Canada. Hiebert, D. (2005) Migration and the Demographic Transformation of Canadian Cities: The Social Geography of Canada's Major Metropolitan Centres in 2017. Working Paper 05-14. Vancouver: Vancouver Centre of Excellence, Research on Immigration and Integration in the Metropolis. Hiebert, D. (2002) "The Spatial Limits of Entrepreneurship: Business Immigrants in Canada" Tijdschrift voor Economische en Sociale Geografie, 93: 173-190. Hiebert, D. (1999) "Immigration and the Changing Social Geography of Greater Vancouver." BC Studies, 121: 35-82. Hiebert, D., P. Mendez, E. Wyly, and V. Preston (Forthcoming) The Housing Situation and Needs of New Immigrants in the Vancouver CMA. Ottawa: Canada Mortgage and Housing Corporation. Hiebert, D., S. D'Addario, K . Sherrell, S. Chan (2005) The Profde of Absolute and Relative Homelessness Among Immigrants, Refugees, and Refugee Claimants in the GVRD. Vancouver: MOSAIC. Last accessed February 1, 2006: http://mosaicbc.com/The_Profile_of_Absolute_and_Relative_Homelessness.pdf Hulchanski, D. (2005) "Rethinking Canada's Housing Affordability Challenge." Discussion Paper, Roundtable on the Canadian Housing Framework Initiative. Last accessed May 25, 2006: www.eichs.org/pdf/AffordableHousing.pdf Hulchanski, D. (2004) " A Tale of Two Canadas: Homeowners Getting Richer, Renters Getting Poorer." In Hulchanski, D. and M . Shapcott (eds.) Finding Room: Policy Options for a Canadian Rental Housing Strategy, 81-88. Toronto: Centre for Urban and Community Studies, University of Toronto. Hulchanski, D. (1995) "The Concept of Housing Affordability: Six Contemporary Uses v of the Housing Expenditure-to-income Ratio." Housing Studies, 10(4): 471-491. 125 Hulchanski, D. and M . Shapcott (eds.) (2004) Finding Room. Policy Options for a Canadian Rental Housing Strategy, 256-261. Toronto: Centre for Urban and Community Studies, University of Toronto. Katz, B., M . Turner, K. Brown, M . Cunningham and N . Sawyer (2003) Rethinking Local Affordable Housing Strategies: Lessons from 70 Years of Policy and Practice. Washington, D . C : Brookings Institute Center on Urban and Metropolitan Policy and The Urban Institute. Last accessed June 27, 2006: http://www.brookings.edu/ es/urban/knight/housingreview.pdf Katz, C. (2001) 'Vagabond Capitalism and the Necessity of Social Reproduction.' Antipode, 33(4): 709-728. Kemeny, J. (1981) " A Critique of Homeownership." In Bratt, R., C. Hartman and A . Meyerson (eds.) (1986) Critical Perspectives on Housing, 272-276. Philadelphia: Temple University Press. Kengott, G. (1912) The Record of a City: A Social Survey of Lowell Massachusetts. New York: Macmillan. Lapointe Consulting Inc. and R. Murdie (1996) Immigrants and the Canadian Housing Market: Living Arrangements, Housing Characteristics and Preferences. Ottawa: Canada Mortgage and Housing Corporation. Laryea, S. (1999) Housing Ownership Patterns of Immigrants in Canada. Working Paper 99-19. Vancouver: Vancouver Centre of Excellence, Research on Immigration and Integration in the Metropolis. Latour, B. (2004) "Why Has Critique Run out of Steam? From Matters of Fact to Matters of Concern." Critical Inquiry, 30: 225-248. Ley, D. (1999) "Myths and Meanings of Immigration and the Metropolis." The Canadian Geographer, 43(1): 2-19. Ley, D. and J. Tutchener (2001) "Immigration, Globalisation and House Prices in Canada's Gateway Cities." Housing Studies, 16(2): 199-223. Ley, D., P. Murphy, K. Olds and B. Randolph (2001) "Immigration and Housing." Progress in Planning, 55: 141-152. Ley, D. and H. Smith (2000) "Relationships Between Deprivation and Immigrant Groups in Large Canadian Cities." Urban Studies, 37(1): 37-62. Ley, D. and H. Smith (1997) "Is There an Immigrant Underclass in Canadian Cities?" Working Paper 97-08. Vancouver: Vancouver Centre of Excellence, Research on Immigration and Integration in the Metropolis. Listokin, D., E. Wyly, B. Schmitt, and I. Voicu (2002) The Potential and Limitations of Mortgage Innovation in Fostering Homeownership in the United States. Washington, D . C : Fannie Mae Foundation. Lo, L. and Wang, S. (1997) "Settlement Patterns of Toronto's Chinese Immigrants: Convergence or Divergence?" Canadian Journal of Regional Science, 20(1,2): 49-72. 126 Mendez, P., D. Hiebert and E. Wyly (2006) "Landing at Home: Insights on Immigration and Metropolitan Housing Markets from the Longitudinal Survey of Immigrants to Canada." Canadian Journal of Urban Research, 15(2): 82-104. Moore, E. and A. Skaburskis (2004) "Canada's Increasing Housing Affordability Burdens." Housing Studies, 19(3): 395-413. Murdie, R. (1994) '"Blacks in Near-Ghettos?' Black Visible Minority Population in Metropolitan Toronto Housing Authority Public Housing Units." Housing Studies, 9(4): 435^157. Murdie, R., A . Chambon, D. Hulchanski and C. Teixeira (1999) "Differential Incorporation and Housing Trajectories of Recent Immigrant Households: Towards a Conceptual Framework." Discussion Paper, Housing New Canadians Working Group, University of Toronto. Last accessed July 2, 2006: http ://www.hnc.utoronto. ca/publish/concept.pdf Murdie, R., V . Preston, M . Chevalier, S. Ghosh (Forthcoming) "Immigrants and Housing: A Review of Canadian Literature From 1990 to 2005." Ottawa: Canada Mortgage Housing Corporation. Murdie, R. and C. Teixeira (2003) "Towards a Comfortable Neighbourhood and Appropriate Housing: Immigrant Experiences in Toronto." In Anisef, P. and M . Lamphier (eds.) The World in a City, 132-191. Toronto: Toronto University Press. Myers, D. and C. Yang Liu (2005) "The Emerging Dominance of Immigrants in the US Housing Market 1970-2000." Urban Policy and Research, 23(3): 347-365. Olds, K. (2001) Globalization and Urban Change: Capital, Culture, and Pacific Rim Mega-Projects. Oxford/New York: Oxford University Press. Orr, S. (2000) A Roof Over Our Heads. Affordable Housing and Urban Growth in Western Canada. Calgary: Canada West Foundation. Owusu. T. (1998) "To Buy or Not to Buy: Determinants of Home Ownership Among Ghanaian Immigrants in Toronto" The Canadian Geographer, 42(1): 40-52. Perin, C. (1977) Everything In Its Place: Social Order and Land Use in America. NJ: Princeton University Press. Pomeroy, S. (2004) "Toward a comprehensive affordable housing strategy for Canada." In Hulchanski, D. and M . Shapcott (eds.) Finding Room: Policy Options for a Canadian Rental Housing Strategy, 237-242. Toronto: Centre for Urban and Community Studies, University of Toronto. Quigley, J. and S. Raphael (2004) "Is Housing Unaffordable? Why Isn't It More Affordable?" Journal of Economic Perspectives, 18(1): 129-152: Quigley, J., S. Raphael, and E. Smolensky (2001) "Homeless in America, Homeless in California." The Review of Economics and Statistics, 83:3 7-51. Ray, B. (1999) "Plural Geographies in Canadian Cities: Interpreting Immigrant Residential Spaces in Toronto and Montreal." Canadian Journal of Regional Science, 22(1,2): 65-86. 127 Ray, B. and E. Moore (1991) "Access to Homeownership Among Immigrant Groups in Canada." Canadian Review of Sociology and Anthropology, 28(1): 1-28. Reid, C. (2004) "Achieving the American Dream? A Longitudinal Analysis of the Homeownership Experiences of Low-Income Households." Unpublished dissertation discussion paper. ReMax (2005) Decade in Review. 1995-2005. Last accessed July 7, 2006: http://www.remax.ca/resources/affordability-report.pdf Retsinas, N . and E. Belsky (2002) "Examining the Unexamined Goal." In Retsinas, N . and E. Belsky (eds.) Low-Income Homeownership. Examining the Unexamined Goal, 1-14. Cambridge, Mass. & Washington, D . C : Joint Center for Housing Studies & The Brookings Institution. Ringheim, K. (1990) At Risk of Homelessness. The Roles of Income and Rent. New York: Praeger. Sassen, S. (2001) The Global City. New York, London, Tokyo. Second Edition. New Jersey: Princeton University Press. Skaburskis, A . (1996) "Race and Tenure in Toronto." Urban Studies, 33(2): 223-252. Smith, C. (2006) "Real Estate Sticker Shock." Georgia Straight, March 9: 47-48. Statistics Canada (2006) Report on the Demographic Situation in Canada. 2003 and 2004. Ottawa: Statistics Canada. Statistics Canada (2003) Spending Patterns in Canada. 2001. Ottawa: Statistics Canada. Steele, M . (1979) The Demand for Housing in Canada. Ottawa: Statistics Canada. Suttor, G. (2004) " A Municipal Perspective on a National Housing Strategy." In Hulchanski, D. and M . Shapcott (eds.) Finding Room. Policy Options for a Canadian Rental Housing Strategy, 256-261. Toronto: Centre for Urban and Community Studies, University of Toronto. TD Economics (2005) "Cooling Housing Market Shows No Sign of Bubble Behaviour." Housing Bubble Watch, July 27. Toronto: TD Bank Financial Group. TD Economics (2003) Affordable Housing in Canada: In Search of a New Paradigm. Toronto: TD Bank Financial Group. Thomas, D. (2005) "Socio-Demographic Factors in the Current Housing Market." Canadian Economic Observer, October: 3.1-3.9. Toronto Board of Trade (2003) Affordable, Available, Achievable. Practical Solutions to Affordable Housing Challenges. Toronto: The Toronto Board of Trade. Last accessed June 24, 2006: http://www.bot.com/assets/staticassets/documents/pdf/ policy/fmalbotaffordablehousereport031 .pdf Urbina, I. (2006) "Keeping It Secret as the Family Car Becomes a Home." The New York Times, April 2: Section 1, p . l . Varady, D. and B. Lipman (1994) "What Are Renters Really Like? Results from a National Survey." Housing Policy Debate, 5(4): 491-531. 128 Walks, A. (2001) "The Social Ecology of the Post-Fordist/Global City? Economic Restructuring and Socio-Spatial Polarisation in the Toronto Urban Region." Urban Studies 38: 407-447. Wyly, E., M . Atia, and D. Hammel (2004) "Has Mortgage Capital Found an Inner City Spatial Fix?" Housing Policy Debate, 15: 623-685. Yates, J. (2002) A Spatial Analysis of Trends in Housing Markets and Changing Patterns of Household Structure and Income. Sydney: Australia Housing and Urban Research Institute. Y i Tong, Z. (2004) Homeownership Affordability in Urban America: Past and Future. Special Report, Fannie Mae Foundation. Last accessed June 16, 2006: http://www.fanniemaefoundation.org/programs/pdf/042904_housing_affordability •pdf 129 A P P E N D I X The stacked bar graphs that make up this Appendix are intended to complement the cross-tabulations included in Chapter 3. It should be noted, though, that the latter were intended to represent qualifying and non-qualifying tenant households as a proportion of all C M A tenant households in a given category or subgroup, while the following graphs represent proportions of all C M A households in a given subgroup. Households with a certain characteristic may for instance display a lower rate of inaccessibility than other households when the rate is expressed in relation to tenant households, but a higher rate than other households when it is expressed in relation to all households. For example, households where Person 1 was aged 55 to 64 in Montreal in 2001 had a higher rate of inaccessibility than the aggregate C M A rate when compared to tenant households only (Table 9), but a lower rate when expressed in relation to all households (Figure 12). For all variables except 'sex of Person 1 and of spouse,' I have provided graphs that represent each sub-category as a proportion of the total C M A universe of households for that variable. The tables provide an indication of the size of every sub-category analysed in this study. 130 Figure 10: Mortgage-qualifying tenant households, by age of Person 1, 2001 Source: Census 2001, P U M F (Individuals file). 131 Figure 11: Mortgage-qualifying tenant households, by household structure, 2001 MONTREAL TOTAL Lone parent families Couples Multi-family households Non-family: 1 person Non-family: 2+ persons TORONTO TOTAL Lone parent families Couples Multi-family households Non-family: 1 person Non-family: 2+ persons VANCOUVER TOTAL Lone parent families Couples Multi-family households Non-family: 1 person Non-family: 2+ persons 0% 20% 40% 60% 80% 100% 0 500,000 1,000,000 H Group 3 • Group 2 • Group 1 • Homeowner households El Total households (all tenure) Source: Census 2001, P U M F (Individuals file). 132 Figure 12: Mortgage-qualifying tenant households, by presence of children in the household, 2001 MONTREAL TOTAL Lone-parent families Couples, w/children Multi-family hhlds, w/children Couples, no children All others, no children TORONTO TOTAL Lone-parent families Couples, w/children Multi-family hhlds, w/children Couples, no children All others, no children VANCOUVER TOTAL Lone-parent families Couples, w/children Multi-family hhlds, w/children Couples, no children All others, no children 0% 20% 40% 60% 80% 100% 0 • Group 3 • Group 2 • Group 1 • Homeowner households 250,000 500, H Homeowner households 000 Source: Census 2001, P U M F (Individuals file). 133 Figure 13: t Mortgage-qualifying tenant households, by marital status of Person 1, 2001 M O N T R E A L T O T A L Married, no children Married, with children Common-law, no children Common-law, w/ children Alt other households T O R O N T O T O T A L Married, no children Married, with children Common-law, no children Common-law, w/ children All other households V A N C O U V E R T O T A L Married, no children Married, with children Common-law, no children Common-law, w/ children All other households nflHHPBfi i.*v*"—i 0% 20% 40% 60% 80% 100% 0 • Group 3 • Group 2 • Group 1 • Total homeowners Source: Census 2001, P U M F (Individuals file). 200,000 400,000 600,000 0 Total households (all tenure) 134 Figure 14: Mortgage-qualifying tenant households, by household size, 2001 MONTREAL TOTAL One person Two persons Three persons Four persons Five persons Six persons or more TORONTO TOTAL One person Two persons Three persons Four persons Five persons Six persons or more VANCOUVER TOTAL One person Two persons Three persons Four persons Five persons Six persons or more 0% 20% 40% 60% 80% 100% 0 150,000 300,000 450,000 El Group 3 • Group 2 H Group 1 • Total homeowners HTotal households (all tenure) Source: Census 2001, P U M F (Individuals file). 135 Figure 15: Mortgage-qualifying tenant households, by highest level of schooling of Person 1, 2001 M O N T R E A L T O T A L Less than high school High school grad Trades cert./diploma Some post-sec. Bachelor and above T O R O N T O T O T A L Less than high school High school grad Trades cert./diploma Some post-sec. Bachelor and above V A N C O U V E R T O T A L Less than high school High school grad Trades cert./diploma Some post-sec. Bachelor and above 0% 20% 40% 60% 80% 100% 0 S G r o u p 3 D G r o u p 2 S G r o u p 1 " T o t a l homeowners 300,000 600,000 900,000 B Total households (all tenure) Source: Census 2001, P U M F (Individuals file). 136 Figure 16: Qualifying tenant households, by language used at home by Person 1, 2001 M O N T R E A L T O T A L Official language spoken at home Non-off. lang. at home, knows Engl ish/French Non-off. lang. at home, no English/French T O R O N T O T O T A L Official language spoken at home Non-off. lang. at home, knows English/French Non-off. lang. at home, no Engl ish/French V A N C O U V E R T O T A L Official language spoken at home Non-off. lang. at home, knows Engl ish/French Non-off. lang. at home, no Engl ish/French *s*»£!iJ»>8,,-7,--', llSiSiiiiiiis 0% 20% 40% 60% 80% 100% 0 650,000 1,300,000 IGroup 3 D G r o u p 2 H G r o u p 1 • Total homeowners EDTotal households (all tenure) Source: Census 2001, P U M F (Individuals file). 137 Figure 17: Mortgage-qualifying tenant households, by citizenship of Person 1, 2001 M O N T R E A L T O T A L Canadian by birth Naturalised, immig. pre-1991 Naturalised, immig. 1991-2001 Not Cdn. , immig. pre-1991 Not Cdn. , immig. 1991-2001 Non-perm, resident T O R O N T O T O T A L Canadian by birth Naturalised, immig. pre-1991 Naturalised, immig. 1991-2001 Not Cdn. , immig. pre-1991 Not Cdn. , immig. 1991-2001 Non-perm, resident V A N C O U V E R T O T A L Canadian by birth Naturalised, immig. pre-1991 Naturalised, immig. 1991-2001 Not Cdn. , immig. pre-1991 Not Cdn. , immig. 1991-2001 Non-perm, resident 2 0 % 4 0 % 8 0 % 1 0 0 % U G r o u p 3 D G r o u p 2 • Group 1 • Total homeowners 0 400,000 800 El Total households 000 1,200,000 (all tenure) Source: Census 2001, P U M F (Individuals file). 138 Figure 18: Mortgage-qualifying tenants, by sex of Person 1 and of spouse (if applicable), 2001 M O N T R E A L T O T A L Male Person 1 or spouse (individuals) Female Person 1 or spouse (individuals) T O R O N T O T O T A L Male Person 1 or spouse (individuals) Female Person 1 or spouse (individuals) V A N C O U V E R T O T A L Male Person 1 or spouse (individuals) Female Person 1 or spouse (individuals) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% El Group 3 H Group 2 D G r o u p 1 • Homeowners Source: Census 2001, P U M F (Individuals file). 139 Figure 19: Mortgage-qualifying tenant households, by visible minority status of Person 1, 2001 Source: Census 2001, P U M F (Individuals file). 140 Figure 20: Tenant households, by ethnic group of Person 1, immigrants only, 2001 (Montreal) T O T A L M O N T R E A L European African Arab West As ian South Asian East and S E Asian Latin, C , S. American Caribbean Multiple ethnicities 0% 20% 40% 60% 80% 100% 0 50,000 100,000 150,000 B Group 3 D G r o u p 2 El Group 1 • Owner households • Total households (all tenure) Source: Census 2001, P U M F (Individuals file). 141 Figure 21: Tenant households, by ethnic group of Person 1, immigrants only, 2001 (Toronto) T O T A L T O R O N T O European African Arab West As ian South Asian East and S E Asian Latin, C , S . American Caribbean Multiple ethnicities 0% 20% 40% 60% 80% 100% 0 100,000 200,000 300,000 400,000 El Group 3 • Group 2 • G r o u p 1 • Owner households 0 Total households (all tenure) Source: Census 2001, P U M F (Individuals file). 142 Figure 22: Tenant households, by ethnic group of Person 1, immigrants only, 2001 (Vancouver) T O T A L V A N C O U V E R European African (N/A) Arab (N/A) West As ian South Asian East and S E Asian Latin, C , S . American Caribbean (N/A) Multiple ethnicities S3J 0% 20% 40% 60% 80% 100% 0 50,000 100,000 . 150,000 • Group 3 D G r o u p 2 • Group 1 • Owner households • Total households (all tenure) Source: Census 2001, P U M F (Individuals file). 143 r Figure 23: Mortgage-qualifying tenant households, by immigration cohort of Person 1, 2001 M O N T R E A L T O T A L Immig. before 1961 Immig. 1961-1970 Immig. 1971-1980 Immig. 1981-1990 Immig. 1991-1995 Immig. 1996-2001 T O R O N T O T O T A L Immig. before 1961 • Immig. 1961-1970 Immig. 1971-1980 Immig. 1981-1990 Immig. 1991-1995 Immig. 1996-2001 V A N C O U V E R T O T A L Immig. before 1961 Immig. 1961-1970 Immig. 1971-1980 Immig. 1981-1990 Immig. 1991-1995 Immig. 1996-2001 I T . * • 0% 20% 40% 60% 80% 100% 50,000 100,000 150,000 200,000 II Group 3 • Group 2 H Group 1 • Total homeowners Source: Census 2001, P U M F (Individuals file). El Total households (all tenure) 144 J Figure 24: Qualifying tenant households, by generational ties to immigration of Person 1, 2001 M O N T R E A L T O T A L 1st generation 2nd generation: both parents born outside Canada 2nd generation: one parent born outside Canada 3rd generation and over T O R O N T O T O T A L 1st generation 2nd generation: both parents born outside Canada 2nd generation: one parent born outside Canada 3rd generation and over V A N C O U V E R T O T A L 1st generation 2nd generation: both parents born outside Canada 2nd generation: one parent born outside Canada 3rd generation and over 0% 20% 40% 60% 80% 100% 0 250,000 500,000 750,000 1,000,0 00 H Group 3 • Group 2 PJ Group 1 • Total homeowners OTota l households (all tenure) Source: Census 2001, P U M F (Individuals file). 145 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092728/manifest

Comment

Related Items