oÉëÉ~êÅÜ=êÉéçêí Overview of the Current State of Knowledge on Societal Outcomes of Housing `je`ÔeçãÉ=íç=`~å~Çá~åë Canada Mortgage and Housing Corporation (CMHC) has been Canada’s national housing agency for more than 60 years. Together with other housing stakeholders, we help ensure that Canada maintains one of the best housing systems in the world. We are committed to helping Canadians access a wide choice of quality, affordable homes, while making vibrant, healthy communities and cities a reality across the country. For more information, visit our website at www.cmhc.ca You can also reach us by phone at 1-800-668-2642 or by fax at 1-800-245-9274. Outside Canada call 613-748-2003 or fax to 613-748-2016. Canada Mortgage and Housing Corporation supports the Government of Canada policy on access to information for people with disabilities. If you wish to obtain this publication in alternative formats, call 1-800-668-2642. Canada Mortgage And Housing Corporation Overview Of The Current State Of Knowledge On Societal Outcomes Of Housing Draft Phase II Research Report Assessing the Societal Outcomes of Housing in Canada: Avenues for Future Research SHS Consulting June 2009 This study was funded by Canada Mortgage and Housing Corporation (CMHC). The contents, views and editorial quality of this report are the responsibility of the authors and CMHC accepts no responsibility for them or any consequences arising from the reader’s use of the information, materials and techniques described herein. This report is provided for general information purposes only. Any reliance or action taken based on the information provided is the responsibility of the user. Readers are advised to consult appropriate professional resources to determine what is suitable in their particular case. CMHC assumes no responsibility for any consequences arising from use of the information provided in the report. research highlight January 2010 Socio-economic Series 10-001 Overview of the Current State of Knowledge on Societal Outcomes of Housing introduction Current Conceptual Thinking The objective of this research project was to assess the current state of knowledge regarding linkages between housing and broader societal outcomes, specifically non-housing outcomes related to education, skills development and employment. The literature review included published research and evaluations of programs that examined any combination of education, skills development and employment-related outcomes of housing. Many authors stressed that economic and social development are driven by housing. Social inclusion or exclusion, economic growth and job creation are all linked to it—housing shapes individual well-being and a broad range of social outcomes because it is closely tied to the qualities of community and social space. Figure 1 illustrates the central role that housing plays in many aspects of a person’s life.2 METHODOLOGY A literature review was conducted to assess the current state of knowledge, identify data gaps, examine related methodological challenges and identify future research challenges. The research aimed to address the following questions: n n n What is the current conceptual thinking on how housing affects societal outcomes related to education, skills development and employment? How can these be measured and what challenges exist? What perspectives exist regarding vulnerable groups with distinct housing needs?1 What are the strengths and weaknesses of the various methodological approaches used to identify, quantify and attribute societal outcomes? One of the key impacts of housing is on health outcomes and well-being, a linkage that has received more attention than perhaps any other. It has been a sub-category of the growing body of work on social inequalities and health. The work done in this area has shown evidence that highlights the positive outcomes that stable housing helps to achieve on the health of individuals.3 Other linkages have also been explored, more specifically the impact that the quality of neighbourhoods can have on non-housing outcomes such as education and employment. Authors have identified geographies of opportunity that can impact the possibilities for residents living in different neighbourhoods. 1 Aboriginal people, seniors, persons with disabilities, youth and children, homeless populations, single-parent households, and recent immigrants and refugees. 2 Carter, Tom and Chesya Polevychok (2004). Housing is Good Social Policy. Canadian Policy Research Networks. 3 Examples of such research include: J.R. Dunn, 2000, Housing and Health Inequalities: Review and Prospects for Research; Dunn, 2002, Housing and Inequalities in Health: A Study of Socioeconomic Dimensions of Housing and Self Reported Health from a Survey of Vancouver Residents; Wilkinson and Marmot, 2003, Social Determinants of Health Second Edition: The Solid Facts; Health Canada, 2007, People, Place and Health; CMHC, 2003, Housing Quality and Children’s Socio-emotional Health. Research Highlight Overview of the Current State of Knowledge on Societal Outcomes of Housing HEALTH CARE EDUCATION Enhancing educational attainment Improving physical and mental health SOCIAL DEVELOPMENT Foundation of family life and social interaction HOUSING COMMUNITY DEVELOPMENT A stabilizing and facilitating role Skills development, investment, capacity building INCOME SECURITY LABOUR FORCE IMMIGRATION Enhancing income security Contributing to stability and mobility Facilitating integration Figure 1 Across a metropolitan region, residents will not have equal opportunities, since employment markets and institutions are not distributed equally across the region. Households often locate in certain areas because of their socio-economic status and are thereby limited to particular employment markets and institutions. Such gaps in socio-economic opportunities between and among neighbourhoods can affect the possible opportunities for those living in a neighbourhood to access both public and private resources. Limitations of Existing Research While many of the reviewed studies did find linkages between housing and education, skills development and employment, the review highlighted the fact that there was often difficulty proving causality. This was partly due to the difficulty of separating family characteristics from location. Another issue that was often stressed, including by the authors of the reviewed studies themselves, was the lack of complete and current data, which can affect the validity of the research. 2 Canada Mortgage and Housing Corporation Although much has been written on housing and the links it has, and can have, to other societal and environmental elements, it is often difficult to determine the strengths and weaknesses of these findings and arguments. One of the goals of the literature review and the report was to evaluate these strengths and weaknesses in order to guide further research into the societal links of housing. FINDINGS Over 100 relevant research studies were examined in conducting this review. Despite a general lack of rigour, there is evidence that housing does have impacts on education, skills development and employment. These findings are not without their caveats, as various studies often found that, for some factors, there was contradictory evidence or simply not enough data or research to conclusively assert a particular outcome. As part of the review, gaps and related methodological challenges were identified, as well as the strengths and weaknesses of the various studies. A summary table of the identified linkages from these studies can be found below. Research Highlight Overview of the Current State of Knowledge on Societal Outcomes of Housing Factors Related to Educational Outcomes The stability of housing has been shown to give rise to improvements in educational performance. Some studies also found positive associations with tenure, with more positive educational outcomes for homeowners, although this is not conclusive in all cases, especially for lower-income owner households. Most positive outcomes seem to stem from the stability that housing offers, with positive outcomes shown in this respect for both owner and renter households. In Canada, a study by Curtis and Phipps (2000) found better educational outcomes for children living in owner-occupied housing; however, a key predictor was found to be the hours of parental time available each week for the children. This time spent with the children significantly improved their successes in school. Housing conditions, such as overcrowding and the quality of the housing, are also associated with education outcomes. Factors like noise, overcrowding and poor housing conditions have been linked to poor educational achievement in children. At the same time, while housing conditions do appear to impact educational performance, much of the current research seems to point to socio-economic status as the predictive factor most related to educational outcomes. Neighbourhood effects are perhaps the most intriguing, as their boundaries are often subjective and can give rise to complex social, economic, political and cultural environments in which housing exists. Nevertheless, a relationship has been documented in the literature between children’s educational attainment and affluence, both at the neighbourhood level and at the individual family level. A Canadian study4 that examined neighbourhood effects on children found that family differences seemed to play a greater role than the quality of the environment. Many authors, however, have cautioned about drawing conclusions between neighbourhood characteristics and family and individual characteristics. While these relationships clearly exist, current findings cannot be considered conclusive until further study is conducted on these linkages. Factors Relating to Skills Development and Employment Some literature found that disincentives to improving employment earnings were inherent in housing programs that require higher rents to be paid as earnings increase. Positive outcomes, however, were highlighted in the Jobs-Plus program5 from the United States, which provided job placement and training assistance and enabled residents to keep more of their earnings. Women in particular gained from being involved in such combinations of housing and employment/skills development programs. Research also showed that social security recipients were more likely to work if located closer to job opportunities. However, socio-economic factors were also shown to mitigate these findings, such as the ability to afford to move to a “better” area with more jobs or the availability and accessibility of means of transportation. The literature review also identified a relationship between housing stability and employment/skills development outcomes, through the increased sense of security of the household. Tenure was not found to have a significant effect on employment and skills development. Vulnerable Groups Although many of these findings concerned a range of different people and groups, the study found a lack of information regarding education, skills development and employment outcomes for those groups deemed most vulnerable (such as single-parent households, Aboriginal people and homeless populations). Some research findings did point to certain outcomes; for example, if women are able to find secure, affordable housing, their capacity to pursue educational and professional goals and their ability to form supportive networks are enhanced. Generally, however, the effects of housing on vulnerable groups are not well documented. 4 Oreopoulos, Phillip (2003). The Long-Run Consequences of Living in a Poor Neighborhood, Quarterly Journal of Economics. 5 Bloom, H.S., Riccio, J.A., Verma, N. (2005). Promoting Work in Public Housing: The Effectiveness of Jobs-Plus. Canada Mortgage and Housing Corporation 3 Research Highlight Overview of the Current State of Knowledge on Societal Outcomes of Housing Issues of Methodology Regarding the impact of housing on education, skills development and employment, the biggest issue that affected the quality and strength of the findings in many cases was that few of the studies used statistical or analytical techniques. In some cases, where empirical methods were used, the age of the data proved problematic. This lack of reliable, accurate and timely data represents a challenge. The most striking knowledge gaps that currently exist are those in the data and information regarding the housing and education, skills development and employment outcomes for vulnerable groups. This represents a large gap in Canadian housing research. In most cases, information exists regarding the needs and the important role of housing for these groups, but very little is available in terms of data and analysis of the linkages themselves. More complete and current knowledge documenting the impacts and outcomes of housing for these individual groups would help strengthen housing policies and programs in Canada for the future. Some studies did use reliable techniques, most significantly the Moving to Opportunity (MTO) and HOPE VI research projects in the United States, which were based on randomized control studies. Many studies were based on other, non-experimental, empirical study methodologies, such as longitudinal and cross-sectional. Other studies involved a mixed-method type, blending literature reviews and empirical analysis. Table 1 summarizes the linkages between housing and education, skills development and employment that were identified in the literature. Table 1 Summary of Identified Linkages6 Education Neighbourhood affluence / educational attainment Family income / educational attainment Stability and transience Tenure (owning versus renting) Housing conditions (physical characteristics: overcrowding, density, physical condition of dwelling, noise levels, and so on). Examples of Indicators Used n Neighbourhood affluence; home ownership rates; residential stability/ instability; achievement scores; literacy of parents; median family income; availability of books in household. n Social development; neighbourhood affluence/poverty; family income; parental education level; neighbourhood family structure. n Family income; number of family moves; number of times children changed schools; tenure type; condition of housing. n n Success at school; family income; labour force participation; tenure; condition of home; civic engagement. Child education performance; condition of housing; tenure; family income; educational services. 6 Conclusions and Strength of Linkage n Some studies have found positive educational outcomes for students and children related to neighbourhood affluence and level of neighbourhood education. n Neighbourhood affluence was noted as having a significant positive effect on reducing dropout rates of female students. n Other studies point to little or no effect of neighbourhood affluence. n Overall, findings are inconclusive: many authors stress that individual and family characteristics cannot be completely discounted, while others highlight incomplete knowledge of these linkages. n Some studies have pointed to higher family educational attainment and income as being associated with educational outcomes. n It was noted that improved educational outcomes arise from housing stability. n Number of moves can also affect likelihood of graduation and educational performance. n Positive association with children’s educational attainment and homeownership was indicated in several studies. n Some studies point to negative aspects, or lack of understanding, of homeownership responsibilities and impacts for low-income households. n Strong evidence showed negative educational impacts from aspects such as noisy homes, overcrowded dwellings, living in poor housing conditions and homelessness. n Studies suggested also that good housing conditions are important for children’s educational outcomes and can have effects (positive or negative) in other areas such as health and employment. See the consultant’s report for a full bibliography of the studies reviewed. 4 Canada Mortgage and Housing Corporation Research Highlight Overview of the Current State of Knowledge on Societal Outcomes of Housing Employment and Skills Development Tenure (owning versus renting) Income and rent Spatial inequalities Neighbourhood affluence / level of education Stability Examples of Indicators Used n Employment and education level. n Education level; skills level; employment earnings. n n n Welfare receipt; neighbourhood poverty rates; job accessibility as jobs per job seeker. Gender of household head; level of education; parental earnings; sibling income comparisons over time; labour market engagement. Rental type (private or public); current employment status; living arrangement; education level; difficulty in paying rent; factors affecting decisions about employment; difficulties in finding employment. Conclusions and Strength of Linkage n There was a lack of evidence and research showing a clear relationship between employment/skills development and tenure. n Tenure is more indicative of socio-economic status and employment status. n Research has found a relationship between income-related rent structures and disincentives to work. Referred to as a “poverty trap” by some, the income-related rent structures penalize tenants for working, especially in lower-paying jobs. n n Review of the Jobs-Plus program in the United States highlighted positive outcomes for participants, especially strong outcomes for women who participated in such endeavours. Housing location can have exclusionary effects on employment: regions where jobs are available can often be accompanied by high house prices, and housing (notably public housing) located in areas of high unemployment can make it more difficult for a person to access work. n Studies have highlighted that people are more apt to work if situated closer to work opportunities. n n A locational factor exists between housing and employment, but authors stress that other factors should also be studied in future work, such as transportation accessibility and concentration of poverty. Studies have found few effects, either positive or negative, of high-poverty or low-poverty neighbourhoods on labour or employment. n Research also points to outcomes being attributable to socio-economic status, and not just moving to “better areas.” n n Other factors were noted, such as mass transit accessibility and concentration of poverty in a neighbourhood. Studies have found positive effects (such as a better sense of security) of stable housing for unemployed tenants, inferring better chances or outcomes in finding employment. n Stability was also noted as having positive effects for women, facilitating support networks and aiding in accessing the workforce. n Research notes that this sense of security is likely an important factor for employment and skills development, through increased self-esteem and economic stability. n More research on this issue would help validate findings in this regard. CONCLUSIONS Overall, the literature is fairly consistent in finding that housing in itself is not the root cause of advantage and disadvantage; rather, it is only one element in a set of interrelated factors that determine the outcome. Some authors contend that improvements in housing are not enough to result in significant improvements in non-housing outcomes. Others point out that, because housing is so connected to the neighbourhood it is situated in, it has a major impact on individual well-being on a broad range of social outcomes. For the individual or household, the evidence points to the observation that, the greater the degree of affordability, security of tenure, choice and quality of accommodation,7 the more positive the impact, that is, the more positive the housing status, the greater the likelihood of positive educational performance, skills development and employment success. 7 Measured in terms of housing characteristics, such as overcrowding, age and state of repair. Canada Mortgage and Housing Corporation 5 Research Highlight Overview of the Current State of Knowledge on Societal Outcomes of Housing CMHC Project Manager: Ed Nera Consultant: Ed Starr, Principal, SHS Consulting Housing Research at CMHC Under Part IX of the National Housing Act, the Government of Canada provides funds to CMHC to conduct research into the social, economic and technical aspects of housing and related fields, and to undertake the publishing and distribution of the results of this research. This fact sheet is one of a series intended to inform you of the nature and scope of CMHC’s research. To find more Research Highlights plus a wide variety of information products, visit our website at www.cmhc.ca or contact: Canada Mortgage and Housing Corporation 700 Montreal Road Ottawa, Ontario K1A 0P7 Phone: Fax: 1-800-668-2642 1-800-245-9274 66751 ©2009, Canada Mortgage and Housing Corporation Printed in Canada Produced by CMHC 11-01-10 Although this information product reflects housing experts’ current knowledge, it is provided for general information purposes only. Any reliance or action taken based on the information, materials and techniques described are the responsibility of the user. Readers are advised to consult appropriate professional resources to determine what is safe and suitable in their particular case. Canada Mortgage and Housing Corporation assumes no responsibility for any consequence arising from use of the information, materials and techniques described. Le point en recherche Janvier 2010 Série socio-économique 10-001 Aperçu de l’état actuel des connaissances sur les retombées sociétales du logement INTRODUCTION Concepts actuels Cette étude visait à évaluer l’état des connaissances sur les liens entre le logement et les retombées sociétales générales, en particulier les retombées autres que celles qui touchent le logement, c’est-à-dire celles qui sont liées à la scolarisation, au perfectionnement des compétences et à l’emploi. Parmi les documents examinés figurent des études et des évaluations publiées portant sur des programmes qui analysent les différentes combinaisons de retombées du logement liées à la scolarisation, au perfectionnement des compétences et à l’emploi. De nombreux auteurs affirment que le logement influe non seulement sur le développement économique et social, mais aussi sur l’inclusion et l’exclusion sociale, la croissance économique et la création d’emplois. Le logement influe sur le bien-être des personnes et sur de nombreuses retombées sociétales du fait qu’il est étroitement lié à la qualité de vie qu’offrent la collectivité et l’espace social. La figure 1 illustre la place centrale qu’occupe le logement dans de nombreuses sphères de la vie d’une personne2. MÉTHODE Un examen de la documentation a permis d’évaluer l’état actuel des connaissances et de déterminer les lacunes statistiques, les problèmes méthodologiques et les éventuels défis en recherche. L’étude avait pour but de répondre aux questions suivantes : n n n À l’heure actuelle, comment conceptualise t-on les effets du logement sur les retombées sociétales liées à la scolarisation, au perfectionnement des compétences et à l’emploi? Comment évaluer l’ampleur de ces retombées et savoir si des difficultés existent? Quelles sont les perspectives des groupes vulnérables dont les besoins en logement sont particuliers1? Quels sont les avantages et les inconvénients des diverses méthodes servant à déterminer, à quantifier et à qualifier les retombées sociétales? L’une des principales incidences du logement, et certainement la plus étudiée de toutes, concerne le lien avec la santé et le bien-être. Ce lien fait l’objet de plus en plus d’études sur les inégalités sociales et la santé. Les résultats de ces études montrent clairement les retombées positives d’un logement stable sur la santé des gens3. D’autres liens ont également été analysés, notamment l’effet de la qualité des quartiers sur les retombées autres que celles qui touchent le logement, comme les retombées liées à la scolarisation et à l’emploi. Des auteurs ont trouvé des facteurs géographiques 1 Peuples autochtones, aînés, personnes handicapées, jeunes et enfants, sans-abri, familles monoparentales, nouveaux immigrants et réfugiés. 2 Tom Carter et Chesya Polevychok, Housing is Good Social Policy, Réseaux canadiens de recherche en politiques publiques, 2004. 3 Citons notamment les études suivantes : J. R. Dunn, Housing and Health Inequalities: Review and Prospects for Research, 2000; Dunn, Housing and Inequalities in Health: A Study of Socioeconomic Dimensions of Housing and Self Reported Health from a Survey of Vancouver Residents, 2002; Wilkinson et Marmot, Social Determinants of Health Second Edition: The Solid Facts, 2003; Santé Canada, Les gens, les lieux, la santé, 2007; SCHL, La qualité du logement et la santé socioémotionnelle des enfants, 2003. AU CŒUR DE L’HABITATION Le Point en recherche Aperçu de l’état actuel des connaissances sur les retombées sociétales du logement SOINS DE SANTÉ SCOLARITÉ Augmentation du niveau de scolarité Amélioration de la santé physique et mentale DÉVELOPPEMENT SOCIAL LOGEMENT Fondement de la famille et des interactions sociales Effet stabilisateur et facilitant SÉCURITÉ DU REVENU MAIN D’ŒUVRE Amélioration de la sécurité du revenu DÉVELOPPEMENT COMMUNAUTAIRE Perfectionnement des compétences, investissement, développement du potentiel IMMIGRATION Élément de stabilité et de mobilité Facteur d’intégration Figure 1 qui peuvent influer sur les occasions qui s’offrent aux habitants de différents quartiers. Par exemple, les habitants d’une même région métropolitaine n’ont pas tous des chances égales d’accéder à l’emploi, car les marchés du travail et les entreprises ne sont pas répartis uniformément dans la région. Les ménages choisissent souvent leur lieu de résidence en fonction de leur situation socio-économique, ce qui les limite à certains marchés de l’emploi et à certaines institutions professionnelles. De tels écarts sur le plan des possibilités économiques au sein d’un même quartier ou entre quartiers peuvent empêcher les habitants d’avoir accès aux ressources publiques comme privées. Limite des études réalisées à ce jour Bien que plusieurs des études révèlent l’existence de liens entre le logement, d’une part, et la scolarisation, le perfectionnement des compétences et l’emploi, d’autre part, notre examen de ces études fait ressortir le fait que les liens de cause à effet sont souvent difficiles à démontrer. Ce problème résulte en partie de la difficulté de séparer les caractéristiques familiales du lieu de résidence. L’absence de données exhaustives et à jour est un autre problème souvent relevé, y compris par les auteurs des études examinées, et mine la validité de la recherche. 2 Société canadienne d’hypothèques et de logement Malgré tout ce qui s’est écrit sur les liens réels et potentiels entre le logement et d’autres aspects sociétaux et environnementaux, il demeure souvent difficile de déterminer quels sont les points forts et les points faibles des constatations et de l’argumentation. L’un des objectifs de l’examen de la documentation et du rapport qui en a résulté était d’évaluer ces forces et ces faiblesses pour orienter d’autres études sur les retombées sociétales du logement. CONstatations Plus de 100 études pertinentes ont été examinées. Malgré un manque de rigueur répandu, il semble que les effets du logement sur la scolarité, le perfectionnement des compétences et l’emploi soient réels. Il faut toutefois prendre ces constatations avec des pincettes. En effet, des études ont souvent révélé que, pour certains facteurs, les faits étaient contradictoires ou que les données ou les recherches étaient tout simplement insuffisantes pour pouvoir tirer des conclusions précises quant aux retombées. L’examen a néanmoins permis de cerner les lacunes et les problèmes méthodologiques, de même que les points forts et les points faibles des diverses études. On trouvera ci-dessous un tableau résumant les liens établis dans le cadre de ces études. Le Point en recherche Aperçu de l’état actuel des connaissances sur les retombées sociétales du logement Facteurs liés à la scolarité Il est prouvé que la stabilité du logement contribue à améliorer les résultats scolaires. Selon certaines études, le mode d’occupation a également des retombées favorables sur les résultats scolaires, surtout pour les propriétaires. Cette observation n’est cependant pas concluante dans tous les cas, notamment dans celui des propriétaires à faible revenu. Les meilleures retombées semblent découler de la stabilité qu’offre le logement, tant pour les propriétaires que pour les locataires. Au Canada, une étude de Curtis et Phipps (2000) indique que le fait pour un enfant d’habiter dans un logement occupé par son propriétaire crée de meilleures retombées sur le plan scolaire. Cependant, le nombre d’heures que consacrent hebdomadairement les parents aux enfants semble être une bonne variable explicative puisque cela améliore significativement la réussite scolaire des enfants. Les conditions de logement, comme le nombre d’occupants et la qualité du logement, influent également sur le rendement scolaire. Des facteurs tels que le bruit, le surpeuplement et les mauvaises conditions de logement sont liés à de mauvais résultats scolaires chez les enfants. Par contre, si les conditions de logement semblent effectivement avoir des effets sur les résultats scolaires, la plupart des recherches menées à ce jour laissent entrevoir que la situation socio-économique est le facteur qui prédit le mieux les résultats scolaires. L’influence des quartiers est probablement le facteur le plus intrigant, car leurs frontières sont souvent subjectives et ils peuvent donner naissance à des environnements sociaux, économiques, politiques et culturels complexes dans lesquels le logement n’est qu’un des éléments. Néanmoins, la documentation étudiée établit un lien entre le niveau de scolarité des enfants et l’aisance du quartier et de la famille. Une étude canadienne4 qui analyse l’influence du quartier sur la situation des enfants conclut que les différences familiales exercent une plus grande influence que la qualité de l’environnement. Cependant, de nombreux auteurs suggèrent de faire preuve de prudence avant de tirer des conclusions de la comparaison entre les caractéristiques du quartier et celles de la famille et des personnes. Bien qu’il soit évident que ces liens existent, les observations actuelles ne pourront être jugées concluantes tant que d’autres études sur ces liens n’auront pas été réalisées. Facteurs liés au perfectionnement des compétences et à l’emploi Selon certaines recherches, les programmes de logement dans le cadre desquels le loyer augmente avec le revenu dissuadent les participants d’augmenter leurs revenus d’emploi. Toutefois, des effets positifs ont été soulignés dans une étude sur le programme américain Jobs-Plus5, un programme d’aide à la recherche d’emploi et à la formation qui a permis aux résidents de conserver une plus grande part de leurs revenus. Les femmes sont les grandes bénéficiaires de ce genre de programme axé à la fois sur le logement, l’emploi et le perfectionnement des compétences. En outre, les études montrent que les bénéficiaires de la sécurité sociale ont davantage de chances de trouver de l’emploi s’ils habitent près des secteurs d’activité économique. Par contre, cette observation est nuancée par des facteurs socio-économiques tels que la possibilité de déménager dans un quartier offrant de meilleures occasions d’emploi ou un accès à des moyens de transport. L’examen de la documentation a aussi permis d’établir un lien entre la stabilité du logement et un meilleur sentiment de sécurité chez les ménages, qui est attribuable aux effets sur l’emploi et le perfectionnement des compétences. Le mode d’occupation ne semble pas avoir d’incidence notable sur l’emploi et le perfectionnement des compétences. Groupes vulnérables Quoique nombre de ces observations concernent des personnes et des groupes différents, l’étude constate un manque d’information sur les incidences de la scolarisation, du perfectionnement des compétences et de l’emploi chez les groupes vulnérables (familles monoparentales, peuples autochtones et sans-abri). Certaines conclusions confirment cependant certaines retombées; par exemple, si les femmes trouvent un logement sûr et abordable, elles renforcent leur 4 Phillip Oreopoulos, The Long-Run Consequences of Living in a Poor Neighborhood, Quarterly Journal of Economics, 2003. 5 H. S. Bloom, J. A. Riccio et N. Verma, Promoting Work in Public Housing: The Effectiveness of Jobs-Plus, 2005. Société canadienne d’hypothèques et de logement 3 Le Point en recherche Aperçu de l’état actuel des connaissances sur les retombées sociétales du logement capacité à atteindre des objectifs liés aux études ou au travail, ainsi qu’à former des réseaux de soutien. Toutefois, les effets du logement sur les groupes vulnérables ne sont généralement pas bien documentés. Problèmes méthodologiques En ce qui concerne les effets du logement sur la scolarité, le perfectionnement des compétences et l’emploi, le principal problème des conclusions tirées, sur les plans de leur qualité et de leur validité, réside dans le fait que peu d’études ont été réalisées à l’aide de méthodes statistiques ou analytiques. Dans certains cas, les études ont été faites selon la méthode empirique et les données utilisées s’avèrent aujourd’hui désuètes. Ce manque de données fiables, exactes et à jour pose problème. À l’heure actuelle, la lacune la plus évidente est l’absence de données et d’information au sujet des retombées en matière de logement, de scolarité, de perfectionnement des compétences et d’emploi chez les groupes vulnérables. Il s’agit d’une importante lacune de la recherche sur le logement au Canada. Dans la majorité des cas, il existe de l’information sur les besoins des groupes vulnérables en matière de logement et sur la place centrale qu’occupe le logement dans leur vie, mais il existe très peu de données et d’analyses sur les liens eux-mêmes. Une documentation exhaustive et à jour sur les effets et les retombées du logement chez ces groupes pourrait contribuer à améliorer les programmes et les politiques du Canada à cet égard. Certaines études ont été menées à l’aide d’une méthodologie fiable, notamment les travaux de recherche américains Moving to Opportunity (MTO) et HOPE IV, fondés sur des études contrôlées sur échantillons aléatoires. Bon nombre d’études reposent sur d’autres méthodes empiriques et non expérimentales, notamment les techniques longitudinales et transversales. D’autres études ont fait appel à une méthode qui combine l’examen de la documentation scientifique et l’analyse empirique. Le tableau 1 résume les liens établis dans la documentation entre le logement et la scolarité, le perfectionnement des compétences et l’emploi. TABLEAU 1 : Résumé des liens établis6 Scolarité Aisance du quartier/ niveau de scolarité Revenu familial/ niveau de scolarité Stabilité et mobilité Mode d’occupation (propriétaire ou locataire) Exemples d’indicateurs n Aisance du quartier; pourcentage de propriétaires-occupants; stabilité/ instabilité du logement; résultats scolaires; alphabétisme des parents; revenu familial moyen; accès à des livres à la maison. n Développement social; aisance/ pauvreté du quartier; revenu familial; niveau de scolarité des parents; structure familiale du quartier n Revenu familial; nombre de déménagements de la famille; nombre de changements d’école; mode d’occupation; état du logement Réussite scolaire; revenu familial; participation au marché du travail; mode d’occupation; état du logement; engagement civique n 6 Conclusions et validité du lien n Certaines études montrent que les élèves et les enfants qui habitent un quartier aisé et alphabétisé obtiennent de meilleurs résultats scolaires que les autres. n L’aisance du quartier contribue grandement à réduire le taux de décrochage scolaire chez les filles. n Selon d’autres études, l’aisance du quartier exerce peu d’influence, voire aucune. n n Globalement, les constatations ne sont pas concluantes : de nombreux auteurs soulignent qu’il faut se garder d’ignorer complètement les caractéristiques individuelles et familiales, alors que d’autres soulignent le manque de connaissances sur ces liens. Certaines études indiquent que les résultats scolaires sont proportionnels au niveau de scolarité et au revenu des membres de la famille. n La stabilité du logement améliore les résultats scolaires. n Le nombre de déménagements peut aussi influer sur la probabilité d’obtention d’un diplôme et les résultats scolaires. n Plusieurs études indiquent que les enfants qui habitent dans un logement occupé par son propriétaire ont un niveau de scolarité plus élevé que les autres. n Certaines études révèlent des aspects négatifs ou un manque de compréhension des responsabilités qui viennent avec le fait d’être propriétaire ainsi que de l’impact sur les ménages à faible revenu. La liste complète des études examinées se trouve dans le rapport du consultant. 4 Société canadienne d’hypothèques et de logement Le Point en recherche Aperçu de l’état actuel des connaissances sur les retombées sociétales du logement Conditions de logement (caractéristiques physiques : surpeuplement, densité, état des lieux, niveau de bruit, etc.) Emploi et perfectionnement des compétences Mode d’occupation (propriétaire ou locataire) Revenu et loyer Inégalités relativement à l’emplacement Aisance du quartier/ niveau de scolarité Stabilité n Résultats scolaires des enfants; état du logement; mode d’occupation; revenu familial; services éducatifs n De nombreux faits concourent à indiquer qu’il y a impact négatif sur les résultats scolaires lorsque, par exemple, un enfant habite dans un logement bruyant, surpeuplé ou en mauvais état, ou s’il est sans abri. n Certaines études semblent aussi indiquer que les conditions de logement ont une forte incidence sur les résultats scolaires des enfants et peuvent influer (favorablement ou non) sur d’autres aspects tels que la santé et l’emploi. Exemples d’indicateurs n n n n n Niveaux d’emploi et de scolarité Niveau de scolarité; niveau de compétence; revenu d’emploi Prestations d’aide sociale; taux de pauvreté du quartier; accès à l’emploi (nombre d’emplois disponibles par chercheur d’emploi) Sexe du chef du ménage; niveau de scolarité; revenu des parents; comparaison du revenu avec celui des frères et sœurs au fil du temps; participation au marché du travail Type de logement locatif (privé ou public); situation d’emploi actuelle; conditions de logement; niveau de scolarité; difficulté à payer le loyer; facteurs influant sur les décisions d’emploi; difficulté à trouver un emploi Conclusions et validité du lien n Peu de faits et de recherches montrent un lien probant entre le mode d’occupation et le perfectionnement des compétences/l’emploi. n n Le mode d’occupation reflète davantage la situation socio-économique et la situation d’emploi. Des études montrent un lien entre une augmentation du loyer en fonction du revenu et la désincitation au travail. Parfois appelé « piège de la pauvreté » par certains, ce type de structure pénalise les locataires parce qu’ils travaillent, surtout quand il s’agit d’emplois mal rémunérés. n L’examen du programme américain Jobs-Plus a permis de constater des retombées favorables pour les participants, surtout pour les femmes. n L’emplacement du logement peut avoir des effets limitatifs sur l’emploi : le prix des logements dans les régions qui offrent de l’emploi est souvent élevé, et il est difficile de trouver un emploi lorsqu’on habite dans un logement (notamment un logement social) situé dans une région où le taux de chômage est élevé. n Des études soulignent que les gens sont plus susceptibles de travailler s’ils habitent près des secteurs d’activité économique. n Un facteur géographique lie effectivement le logement à l’emploi, mais les auteurs soutiennent que d’autres facteurs doivent être étudiés, comme l’accès au transport et la concentration de la pauvreté. n Les études révèlent peu d’effets (positifs ou non) exercés par les quartiers à forte ou à faible pauvreté sur la main-d’œuvre ou l’emploi. n La recherche laisse aussi entrevoir que les retombées seraient attribuables non seulement au déménagement dans un « meilleur quartier », mais aussi à la situation socio-économique. n n Parmi les autres facteurs constatés, citons l’accès au transport en commun et la concentration de la pauvreté dans un quartier. Des études ont constaté des effets positifs de la stabilité du logement pour les locataires sans emploi (comme l’augmentation du sentiment de sécurité), ce qui laisse croire à de meilleures chances de trouver un emploi. n La stabilité a aussi une incidence favorable sur la situation des femmes, car elle leur permet de créer des réseaux de soutien et d’accéder à l’emploi. n Selon les études, le sentiment de sécurité est probablement un facteur important pour l’emploi et le perfectionnement des compétences, car il augmente l’estime de soi et la stabilité économique. n D’autres études sur le sujet permettraient de valider les conclusions. CONCLUSION Une constatation générale se dégage de l’examen de la documentation : le logement n’est pas à la source des avantages et des inconvénients; il n’est qu’un des éléments dans un ensemble de facteurs interdépendants. Certains auteurs soutiennent que l’amélioration des conditions de logement ne peut à elle seule améliorer significativement les retombées non liées au logement. D’autres font remarquer que, parce qu’il est étroitement lié au quartier, le logement influe considérablement sur le bien-être des personnes en ce qui a trait à de nombreuses retombées sociétales. Du point de vue du particulier ou du ménage, les faits corroborent l’observation suivante : l’abordabilité des loyers, la sécurité d’occupation, le choix et la qualité du logement7 sont directement proportionnels aux effets engendrés. Ainsi, plus la situation de logement est bonne, meilleures sont les chances d’obtenir de bons résultats scolaires, de perfectionner ses compétences et de trouver un emploi. 7 Mesurés selon les caractéristiques du logement, comme le surpeuplement, l’âge et l’état. Société canadienne d’hypothèques et de logement 5 Le Point en recherche Aperçu de l’état actuel des connaissances sur les retombées sociétales du logement Directeur de projet à la SCHL : Ed Nera Consultant pour le projet de recherche: Ed Starr, Principal, SHS Consulting Recherche sur le logement à la SCHL Aux termes de la partie IX de la Loi nationale sur l’habitation, le gouvernement du Canada verse des fonds à la SCHL afin de lui permettre de faire de la recherche sur les aspects socio-économiques et techniques du logement et des domaines connexes, et d’en publier et d’en diffuser les résultats. Le présent feuillet documentaire fait partie d’une série visant à vous informer sur la nature et la portée du programme de recherche de la SCHL. Pour consulter d’autres feuillets Le Point en recherche et pour prendre connaissance d’un large éventail de produits d’information, visitez notre site Web au www.schl.ca ou communiquez avec la Société canadienne d’hypothèques et de logement 700, chemin de Montréal Ottawa (Ontario) K1A 0P7 Téléphone : 1-800-668-2642 Télécopieur : 1-800-245-9274 66752 ©2009, Société canadienne d’hypothèques et de logement Imprimé au Canada Réalisation : SCHL 13-01-10 Bien que ce produit d’information se fonde sur les connaissances actuelles des experts en habitation, il n’a pour but que d’offrir des renseignements d’ordre général. Les lecteurs assument la responsabilité des mesures ou décisions prises sur la foi des renseignements contenus dans le présent ouvrage. Il revient aux lecteurs de consulter les ressources documentaires pertinentes et les spécialistes du domaine concerné afin de déterminer si, dans leur cas, les renseignements, les matériaux et les techniques sont sécuritaires et conviennent à leurs besoins. La Société canadienne d’hypothèques et de logement se dégage de toute responsabilité relativement aux conséquences résultant de l’utilisation des renseignements, des matériaux et des techniques contenus dans le présent ouvrage. National Office 700 Montreal Road Ottawa ON K1A 0P7 Telephone: (613) 748-2000 Bureau national 700 chemin de Montréal Ottawa ON K1A 0P7 Téléphone : (613) 748-2000 Puisqu’on prévoit une demande restreinte pour ce document de recherche, seul le résumé a été traduit. La SCHL fera traduire le document si la demande le justifie. Pour nous aider à déterminer si la demande justifie que ce rapport soit traduit en français, veuillez remplir la partie ci-dessous et la retourner à l’adresse suivante : Centre canadien de documentation sur l’habitation Société canadienne d’hypothèques et de logement 700, chemin Montréal, bureau C1-200 Ottawa (Ontario) K1A 0P7 Titre du rapport: _______________________________________ _______________________________________ Je préférerais que ce rapport soit disponible en français. NOM _____________________________________________ ADRESSE___________________________________________ rue App. ___________________________________________________________ ville province Code postal No de téléphone ( ) ____________ Table of Contents 1. Introduction ....................................................................................................1 1.1 Research Objectives ..................................................................................2 2. Current Conceptual Thinking on How Housing Affects Societal Outcomes Related to Education, Skills Development and Employment: Findings of the Literature Review .....................................................................5 2.1 Background and Context ............................................................................7 2.2 How Housing May Affect Societal Outcomes Related to Education, Skills Development and Employment .........................................................................8 2.2.1 Housing Factors Related to Educational Outcomes ............................9 • Neighbourhoods................................................................................10 • Transience and Stability....................................................................11 • Tenure...............................................................................................12 • Housing Conditions ...........................................................................12 2.2.2 Factors Relating to Skills Development and Employment..................13 • Tenure...............................................................................................13 • Income and Rent Structure ...............................................................14 • Location ............................................................................................14 • Stability .............................................................................................15 2.3 Impact of Housing on Vulnerable Groups.................................................15 • Aboriginal Peoples ............................................................................17 • Women and Single-Parent Households ............................................18 • Children and Youth ...........................................................................19 • Seniors..............................................................................................20 Recent Immigrants and Refugees.....................................................21 • • Housing and Homeless Populations .................................................22 2.4 Summary of Identified Linkages ...............................................................25 3. Methodological Approaches .......................................................................29 3.1 Methodologies and Data ..........................................................................29 3.1.1 Methodologies Reviewed ..............................................................29 • Non-Experimental Studies ................................................................29 • Experimental Studies ........................................................................30 3.1.2 Available Methodologies and Data Sets: Challenges of Attributing Outcomes to Housing ..................................................................................31 3.2 What Data and Literature Gaps Exist? .....................................................35 4. Addressing Weaknesses and Filling the Gaps: Future Avenues ............38 4.1 Strengthening Methodologies and Data Collection...................................38 4.1.1 Defining the Issues ...........................................................................38 4.1.2 Analytical Models...........................................................................41 Experimental Models................................................................................41 Mixed-Method Models ..............................................................................43 Longitudinal Models .................................................................................44 Other Models............................................................................................45 4.1.3 Data and Data Sets .......................................................................46 Longitudinal data sets ..............................................................................47 The Longitudinal Administrative Database (LAD).....................................47 Cross-section data sets............................................................................47 The Longitudinal Survey of Immigrants to Canada (LSIC) .......................48 Methodological Notes of Caution .............................................................49 4.1.4 Considerations for Indicators of Housing and Education, Skills Development and Employment Impacts ......................................................50 4.2 Future Research Considerations..............................................................54 5. Concluding Discussion ...............................................................................57 Appendices .......................................................................................................58 Appendix I - Proposals for Future Research Projects ...................................59 Appendix II – Review of Identified Methodologically Sound Studies...........61 List of Acronyms...............................................................................................73 Bibliography ......................................................................................................74 1. Introduction This study is being undertaken to help increase the level of understanding regarding the linkages between housing and broader societal outcomes, specifically education, skills development, and employment. In particular, the project is aimed at identifying and assessing the current body of research regarding these linkages in terms of existing empirical evidence, data gaps, and related methodological challenges and opportunities. For example, a recent study by Habitat for Humanity and the University of Guelph titled “Assessment of the Outcomes for Habitat for Humanity Home Buyers” (2004) provides an analysis of 454 Habitat homes from across Canada. The study identifies several benefits for the residents due to their improved housing situation. In particular, the report states that almost one quarter (23.6%) of the household respondents indicated that one or both spouses had returned to school following moving into their new home. Further, almost half of these individuals learned a new trade or upgraded their job skills. In addition, the report states that approximately 30% of the individuals who returned to school enrolled in a college or university program. Respondents indicated that the stability of their housing costs allowed them the freedom to consider returning to school. A study on job creation and housing construction (Saks, 2005) prepared as part of the Federal Reserve Board Finance and Economics Discussion Series (Washington DC) explores the impact of housing supply regulations on housing and local labour markets. The study determines that housing regulations can change the geographic distribution of housing prices and alter the pattern of labour migration. As a result, the report finds that employment growth is lower in places where the housing supply is more constrained. The study discusses this finding to assess the impact of these policies on individuals of low and moderate income. These types of research point to the linkages between housing and other key societal outcomes such as education, skills development and employment. By learning more about these linkages, CMHC and other housing policy agencies can have a greater understanding of the broader impacts of various housing policy initiatives on society as a whole. This study is being undertaken to investigate the available research on such impacts, identify the strength of any linkages that have been determined and identify any research gaps that should be filled to improve our understanding of the subject. 1 1.1 Research Objectives The first phase of the research carried out for this project consisted of a review and assessment of literature pertaining to the linkages between housing and other broader societal outcomes (SHS, 2007). Based on this review, a working definition of societal outcomes of housing related to education, skills development and employment was developed, as follows: Housing conditions and characteristics in Canada are linked to the level of educational performance, skills development and employment opportunities of the individual/household and society as a whole; that is, the more positive the housing situation in terms of affordability, security of tenure, choice and quality of accommodation, the more positive the impact on education, skills development and employment. While many of the studies did find linkages between housing and education, skills development and employment, there was often difficulty proving causality. One reason for this is the difficulty of separating family characteristics from location; another problem is the lack of complete and current data: in many cases, these weaknesses are particularly problematic for the validity of the research. Nevertheless, much of the research has established, despite a general lack of rigour, that there is evidence that housing does have impacts on education, skills development and employment. This review identified gaps and related methodological challenges, as well as strengths and weaknesses of the various studies. Building on the findings of the literature review, this paper focuses on answering the following research questions: • What is the current conceptual thinking on how housing affects societal outcomes related to education, skills development and employment? Evaluate how these effects may be measured and any challenges that accompany this measurement. Additionally, what are the perspectives of such studies with respect to different vulnerable groups with distinct housing needs, such as seniors, persons with disabilities, youth and children, homeless populations, Aboriginal peoples, women and single-parent households, and recent immigrants and refugees. • What were the strengths and weaknesses of the various methodological approaches that have been used to identify, quantify and attribute societal outcomes between housing and education, skills development and employment? Specifically, attention will be 2 given to what challenges existed, or exist and how they were overcome, as well as what data gaps exist in the current knowledge. • What future research, including programs and individual studies, could be recommended as feasible and most promising? Given the current knowledge of the linkages between housing and employment, skills development and education, how might the approaches best be put to use, how could data gaps best be filled, and what indicators exist or could be created to measure societal outcomes of housing. Addressing these research questions is of particular importance in starting to establish a more formal set of practices and methodologies to more accurately and regularly measure the societal outcomes of housing. This would enable governments, researchers, and other housing stakeholders to develop and further improve upon policies and programs based on proven and valid research findings. In undertaking the review of research conducted on the linkages between housing and education, skills development and employment, it became clear that some key concepts must be kept in mind in any future work in this area. That is, the research has found that housing is situated geographically in neighbourhoods, which may be rural, urban, or suburban, affluent or poor, large or small, racially and ethnically diverse or homogeneous. In order to discuss the impact of housing on non-housing outcomes, housing needs to be viewed in the context of both the neighbourhood - the physical area around the house and home - and the community, which includes the social characteristics and the services provided in the neighbourhood. Housing is thus a multi-dimensional element, and can be seen from many different yet complementary angles, which include housing “as a physical commodity, as an economic entity, as an item tied to different forms of consumption, and as a location in space” (Shlay, 1995, p.699). The concept of neighbourhood is more difficult to define than that of housing. In a recent study from the University of Toronto that examined income polarization in that city’s neighbourhoods, David Hulchanski highlighted the fact that defining them is an inherently subjective process, as “there is no way to draw boundaries that define specific neighbourhoods” (Hulchanski, 2007, p.3). He further stresses that this is because neighbourhoods encompass many aspects and elements of each resident’s sense of community and experiences. The study did stress that regardless of the difficulty in defining what exactly constitutes neighbourhoods, there is no doubt about their impacts on health status, education outcomes and other aspects of personal well-being. 3 When considering the effects of housing on education, skills development and employment, therefore, it becomes clear that these linkages between the elements also encompass linkages to other aspects, such as neighbourhoods. 4 2. Current Conceptual Thinking on How Housing Affects Societal Outcomes Related to Education, Skills Development and Employment: Findings of the Literature Review In The Truly Disadvantaged, written in 1987, Wilson put forth the argument that there are strong neighbourhood effects on the behaviour and social outcomes of residents. Wilson posited that when black middle class families moved from inner-city areas in the 1960’s, they left the poor of those neighbourhoods socially disadvantaged. As a result of the loss of economically stable families, schools and local businesses in these areas could not be sustained during the 1970s and 1980s periods of high unemployment. Wilson believed that the lack of educational opportunities, role models and social capital resulted in the poor of these neighbourhoods becoming dependent on welfare and entrenched in poverty. Many authors have stressed that economic and social development are driven by the housing system. Social inclusion or exclusion, economic growth and job creation are all linked to it: “the housing system shapes individual well-being and a broad range of social outcomes because it is closely tied to the qualities of community and social space” (Jackson, 2004). A paper by Andrew Jackson cites the TD Bank (2003), Fallis (1994) and Maclennan (2001), all of whom identify the important role housing plays in both economic and social development. Maclennan’s work makes the case that both social and economic progress are shaped, in large part, by housing since economic development is socially grounded and it evolves in place. This is echoed in Carter and Polevychok’s (2004) research where they reiterate that homes do not exist in isolation in their neighbourhoods. They identify the need for community-level interventions that consider housing as a major factor related to broader community-based initiatives because of the combined influence on people of the intersection of housing and neighbourhood. Hay (2005) uses a figure from Carter and Polevychok (Figure 1) to illustrate the links between housing and its role in securing the well-being of the population as a whole (p. 2). It presents the central role that housing plays in many aspects of a person’s life and even though the focus of the present study is education, skills development and employment, the importance of housing becomes clear. 5 Figure 1 HEALTH CARE EDUCATION Enhancing educational attainment Improving physical & mental health SOCIAL DEVELOPMENT Foundation of family life & social interaction INCOME SECURITY Enhancing income security HOUSING A stabilizing and facilitating role LABOUR FORCE Contributing to stability and mobility COMMUNITY DEVELOPMENT Skills development, investment, capacity building. IMMIGRATION Facilitating integration Source: Carter & Polevychok, 2004 As noted in Figure 1, one of the key impacts of housing is on health care. This linkage has received more attention than perhaps any other, having seemingly been a sub-category of the growing body of work on social inequalities and health (Dunn 2000). The work done in this area has shown evidence highlighting the positive outcomes that stable housing helps achieve on the health of individuals.1 1 Examples of such research are not exclusive to, but include the following: J.R. Dunn, 2000, Housing and Health Inequalities: Review and Prospects for Research; Dunn, 2002, Housing and Inequalities in Health: A Study of Socioeconomic Dimensions of Housing and Self Reported Health from a Survey of Vancouver Residents; Wilkinson & Marmot,2003, Social Determinants of Health Second Edition: The Solid Facts; Health Canada, 2007, People, Place and Health; CMHC, 2003, Housing Quality and Children’s Socioemotional Health. 6 Galster and Killen (1995) also emphasized such linkages, more specifically the impact that the quality of neighbourhoods can have on the development of nonhousing outcomes such as education and employment. They identified “a geography of metropolitan opportunity” (p. 8), which impacts the possibilities for residents living in different neighbourhoods: across a metropolitan region, residents will not have equal opportunities since markets and institutions are not distributed equally across the region. This means that households are often confined to certain areas because of their socio-economic status (SES) and are therefore limited to particular markets and institutions. Such gaps in socioeconomic opportunities between and among neighbourhoods can affect the possible opportunities for those living in a neighbourhood in terms of both public and private resources. While it is clear that much has been written on housing and the links it has, and can have, to other societal and environmental elements, it is often difficult to determine the strengths and weaknesses of these findings and arguments. One goal of the literature review was to evaluate these strengths and weaknesses, in order to guide further research into the societal links of housing. The following section addresses the various outcomes that were drawn from the Phase One Literature Review. 2.1 Background and Context During the literature review, it became clear that the recent interest in examining the societal outcomes of housing stems largely from three important programs in the United-States. Some aspects of education, skills development and employment outcomes were examined as part of the Gautreaux Program in Chicago that involved over 7000 families between 1976 and 1998. This program came about as the result of a 1976 Supreme Court decision in a lawsuit filed by public housing residents against the U.S. Department of Housing and Urban Development (HUD) and the Chicago Housing Authority, charging them with racially discriminatory policies (Grant, 2001). Public housing residents, including those on the waiting list, were allowed to receive vouchers (Section 8 housing certificates) and move to private-sector apartments as a deliberate attempt to create racial and social class mixing. These apartments were located either in the mostly-white suburbs or in the City of Chicago. Families were not allowed to choose where in the Chicago area they would be placed and were guided by placement counsellors who helped them review the advantages and disadvantages of moving, and took them to visit the apartments and the communities. About half of the 7000 families moved to the suburbs (Rosenbaum, Reynolds, & Deluca, 2001). The positive non-housing social and economic outcomes resulting from this longitudinal study were instrumental in later developments in housing projects undertaken in the United States. Studies found that children 7 who had moved to the suburbs grew to be young adults who were much more likely to graduate from high school, attend four-year colleges, and have jobs with better pay and benefits (Rosenbaum, Reynolds & DeLuca, 2001). Gautreaux mothers who moved to the suburbs had higher employment rates than city movers. As a result of the Gautreaux program, a five-city national demonstration program called Moving to Opportunity (MTO) was established in 1991. Approximately 4600 families in Baltimore, Boston, Chicago, Los Angeles and New York City have participated in this project where participant eligibility was limited to very low-income families with children who lived in public housing or Section 8 (voucher subsidized) project-based housing located in central city neighbourhoods with high concentrations of poverty.2 In addition, in 1992, HOPE VI was established as the primary U.S. federal program to reform public housing. HOPE VI is a program that replaces the most distressed public housing units with mixed income housing. Both MTO and HOPE VI were designed to create mixed-income communities. The MTO model provides housing vouchers to low income families that can be used to help pay market rent only in low poverty neighbourhoods and HOPE VI helps relocate lowincome families to mixed income buildings or developments. Research related to the non-shelter outcomes of these housing models is ongoing. “If, as a result of moving from concentrated ‘ghettos’ of poverty to mixed-income communities, poor workers and families are able to find jobs that allow for advancement beyond minimum wage, find out about and enrol in training programs, send their children to better schools, and even complete post secondary education, rates of poverty and inequality can be expected to improve” (Grant, 2001). 2.2 How Housing May Affect Societal Outcomes Related to Education, Skills Development and Employment Overall, the literature is fairly consistent in finding that housing in itself is not the root cause of disadvantage; rather it is only one element in a set of interrelated factors that determine advantage and disadvantage. Some authors point out that improvements in housing are not enough to result in significant improvements in non-housing outcomes (Mullins & Western, 2001a). Others, such as Jackson (2004), point out that because housing is so connected to the neighbourhood it is situated in, it has a major impact on individual well-being on a broad range of social outcomes. The latter also identified the link between socio-economic status, child outcomes, and housing since high housing costs push families 2 Information based on the MTO interim evaluation. The current findings have been evaluated against the findings of other studies in this report. 8 deeper into poverty, often resulting in families living in deprived neighbourhoods that lack social supports and social services, have high rates of crime, and poorquality schools. For the individual or household, the evidence points to the observation that the greater the degree of affordability, security of tenure, choice, and quality of accommodation3 the more positive the impact: i.e. the more positive the housing status, the greater the likelihood of positive educational performance, skills development and employment success. Pomeroy (2004) states that, “it is the role housing plays in enabling citizens to fully participate in society, or not, that needs to be made central to the policy and program debate” (p. 18). The Phase One Literature Review(SHS, 2007), which assessed the literature, established that, despite a general lack of rigour in much of the research work conducted on the subject, there is evidence that housing has positive impacts on education, skills development and employment. When talking of housing it is important to recognize that housing refers not only to the physical structure of the dwelling, including its design and characteristics, but also to the social and psychological aspects of the house, which might be referred to as “the home”. “Home is a base that is integral to people’s emotional, cultural, social and economic health” (Carter & Polevychok, 2004). Nevertheless, many of the studies did highlight significant links between housing and the three elements that are employment, skills development and education. This is of particular importance within the mandate of this research, supporting the need for further research on these issues. The importance of research in these areas can be seen in the amount of work done surrounding the MTO and HOPE IV programs: the findings to date, drawn from the various studies, have shown interesting links between housing and other elements, as well as the impacts of housing. Such evidence can make a valuable contribution to the development of housing policies and programs. It should also be noted that the MTO studies to date have been using data from the interim evaluation. The program is currently approaching the final data collection process, and therefore judgement as to the ultimate impacts, and strengths of the linkages should be reserved for the final evaluations. 2.2.1 Housing Factors Related to Educational Outcomes Within the literature reviewed, several housing-related factors were identified as having, or potentially having, an impact on educational outcomes. These elements were the following: • • 3 Neighbourhoods Transience and stability Measured in terms of housing characteristics such as overcrowding, age, state of repair, etc. 9 • • Tenure type Housing conditions (physical characteristics of the house such as overcrowding and density) Their impact on educational outcomes is summarized below. • Neighbourhoods It would seem that “neighbourhood” is a rather large and somewhat subjective term. This is also compounded by much of the findings of the research that highlights the cross-cutting nature of many of the elements in question that are at play within a neighbourhood. For example, one cannot necessarily evaluate neighbourhood effects without considering housing characteristics or quality, or human factors. Many authors highlight this, cautioning to try and understand the interplay between characteristics of the neighbourhood, the family and individuals, and the community(ies). This can be taken one step further by asking how one separates all of these effects from that of the housing units in which people live. This is not to say the above elements cannot be assessed individually, but one must be aware of the inherent complexity and the potential for cumulative effects that larger socio-economic structures such as neighbourhoods can have on research findings, making it difficult to evaluate each component. However, housing itself is affected by the environment around it, thus it is important to consider this when evaluating the impacts of housing outcomes on individuals and families. Regardless of the issues arising from its definition, an insightful observation by Ellen and Turner highlights the general findings of their work on neighbourhoods and their impact on educational outcomes: “Empirical research generally confirms that neighbourhood environment has an influence on important educational outcomes for children and adults. But efforts to identify which neighbourhood characteristics matter most, and to quantify their importance for families and children, have been inconclusive overall” (Ellen & Turner, 1997, p. 833). This is echoed by many other research studies such as Beauvais & Jenson (2003), Buck (2001), Erebus (2005), Jencks & Mayer (1990), and Manski (1993). One neighbourhood factor that has been shown to make a difference in students’ educational attainment is affluence, highlighted in work by Edwards (2005), Pebley and Sastry (2003) and Kohen & Hertzman (1999 – with Brooks-Gunn in 1998). This was also highlighted in a study of the educational achievements of grade 3 students in Ontario by Tremblay et al. (2001). The research highlighted a correlation between the affluence and educational attainment of a neighbourhood, and higher outcomes for the schools in these neighbourhoods, compared to less-advantaged neighbourhoods. However, Tremblay et al. stress 10 the importance of student characteristics, as well as individual familial characteristics and dynamics can also influence outcomes. Some studies and their authors have shown that higher income and higher levels of education within a family are also associated with higher levels of children’s competencies in school (Edwards, 2005; Peck, 2001; Kohen & Hertzman, 1998). Mullins and Western (2001b) noted that lower income households and those in public housing were less likely to have higher rates of education, but cautioned that their study was cross-sectional, and therefore not necessarily able to identify the positive or negative aspects of living in public housing something better examined through more detailed longitudinal studies. Finally, other studies undertaken to look at the impact of educational attainment on students living in public housing found little differences between their academic achievement and the academic achievement of students whose families used vouchers to move to housing in low-poverty areas (Currie & Yelowitz, 1999; Jacob, 2003; Phipps & Young, 2005; Sanbonmatsu et al,, 2004; U.S. Dept. of Housing and Urban Development, 2003). However, differences in school dropout rates were found between the experimental and control groups for female students living in mixed housing through the use of vouchers: the female students had lower dropout rates when they lived in low poverty areas (Kling & Liebman, 2004, Popkin, Leventhal & Weismann, 2006). Thus, based on these findings, neighbourhood effects do exist and cannot be discounted. However these cannot be entirely relied upon until more conclusive links to overall neighbourhood effects can be established and better understood. Buck (2001) stresses caution in treating these associations and effects given the complexity of issues and elements that form a neighbourhood. • Transience and Stability According to Phibbs and Young (2005), teachers identified stability as important for the learning process, especially when children have learning difficulties. Their findings show that, where public housing is provided, improvements in educational outcomes arise as a result of stability. Poor school attendance and a change of school are both elements that can result in educational problems, often preventing a continuity of assistance or remediation. Changing schools was often associated with poorer educational outcomes (Braconi, 2001; Kohen, Hertzman and Brooks-Gunn, 1998; Wood, et al, as cited in Cooper, 2001). The study on Housing and Schooling conducted by Braconi (2001) for the Citizens Housing and Planning Council (CHPC) concluded that, as the number of moves increase during an adolescent’s school years, the likelihood of graduation is somewhat reduced. Braconi analyzed several housing variables and in general found them to have significant effects on high school completion. He 11 reported finding a positive and statistically significant effect of home ownership on graduation rates, which was consistent with other studies. • Tenure A positive association between home ownership and improvement in children’s educational attainment has been established by several studies (Braconi, 2001; Bridge, et al., 2007; Crawford and Londerville, 2004; Curtis & Phipps, 2006; Mullins & Western, 2001a; Rossi & Weber, 1996; Steele & Sarker, 2005). However, in many cases tenure and income, or socio-economic status, are all closely linked, as owning is more expensive than renting. Many lower income households are therefore less likely to own their dwellings, or rely on public housing units. According to Blunden (2005), “tenure type can be read as a ‘proxy’ for income, employment history, education, etc. not because the tenure type ‘causes’ a particular outcome” (p. 13). Overall it would seem that SES is a more important indicator that is predictive of educational outcomes (Peck, 2001). The analysis of the second wave of data from the National Longitudinal Survey of Children and Youth (NLSCY) done by Curtis and Phipps (2000) examined economic factors in child development. Among these, they measured for housing owned by the family of the child, as well as the quality of the home in question and controlled for assets available to the family, and time associated with acquiring income. They found that homeownership was associated with better outcomes for children’s overall health. However, specifically regarding educational outcomes, their study found that available parent time (whether two parents or a lone parent) was more significantly associated to success in school than housing variables. Therefore, although the association with tenure type has been shown to be positive in some cases, there is evidence that suggests that SES and the stabilizing effect of housing may be more important factors to positive educational achievement than tenure type. Examples of this are seen in the work of Phibbs and Young, highlighting the positive effects of stability from public housing, or Curtis and Phipps where parental time also played an important role. When considering the findings of the MTO and HOPE VI research this would seem to be true. However, caution must be exercised when considering these elements as, again, the extent of the inter-relationships is not fully understood, and it would seem that although tenure type is a factor in educational outcomes, these are influenced by other factors as well. • Housing Conditions In most studies, housing conditions that were studied included the physical characteristics of the dwelling, as well as elements such as overcrowding. Several of the studies reviewed found that overcrowding, noise, and poor 12 housing conditions are linked to educational achievement (Cooper, 2001; Mullins, Western, & Broadbent, 2001b; Phibbs, Young, 2005; Curtis and Phipps, 2000). Mullins and Western (2001a) also review the literature related to housing and education and find that negative education impacts on children occur when they are homeless, or live in noisy homes, overcrowded conditions, or in slums. This is echoed by Curtis and Phipps who found a relationship in the Canadian data they used, showing negative outcomes for children living in homes in need of major repairs. In addition, Braconi’s 2001 study also found that there was some causality with respect to housing quality and educational attainment, relating the number of deficient maintenance conditions present within the unit to the probability of graduating high school, i.e., probability drops by 1% with each additional maintenance condition. However, they suggest further study is required on the connection between substandard housing and educational achievement. 2.2.2 Factors Relating to Skills Development and Employment Within the literature, several housing-related factors were identified as having, or potentially having, an impact on skills development and employment. These elements were the following: • • • • Tenure type Income and Rent Structure Location Stability Their impact on skills development and employment is summarized below. • Tenure Linking tenure to employment and skills development tends to be problematic based on current research. According to Blunden, people’s resources and income result in them sorting themselves into different types of housing. She asserts that “tenure is a de facto indication of employment status and work history” (Blunden, 2005), but follows up by stressing that it is indicative of employment status rather than causative. This is supported by Mullins and Westerns’ (2001b) analysis of Australian data, which showed a significant correlation between education and employment status. Interestingly, however, public housing tenants and low-income private tenants had the lowest scores of all the tenure groups (including owneroccupiers, purchasers, and other private tenants), yet their multiple regression analyses did not show a correlation between these factors and housing per se. 13 Again, these varying conclusions illustrate the need for more complete research and data on these factors. Mullins and Western themselves caution the readers in the interpretation of their data since their analysis was cross-sectional - a “snapshot in time” of sorts – but that better knowledge of what longer terms effects can come of these correlations, and the role housing plays in the equation could only be obtained through the use of longitudinal studies. • Income and Rent Structure Fixed rent rules and rents are often a disincentive for tenants in public housing to find employment (Hulse & Randolph, 2004, Riccio, 2007). This is due to the income-related rent structure in public housing, which creates a very real disincentive for some renters to work: Riccio’s work cites HUD data showing that for roughly every dollar of an income increase, residents’ rents increased accordingly and their housing subsidy decreased. Phibbs and Young (2005) also highlight this “poverty trap” that public housing renters can fall into when they see themselves as being penalized for taking on even part time or casual work. The Jobs-Plus program in the United-States provides assistance with job placement and training, new rent rules where working residents can keep more of their earnings, and a community support for work component. It has been shown to be successful (Bloom, Riccio, Verma, 2005; Kramer, 2000). Johnson and Ruddock (2000) identified a number of positive outcomes for women when they were involved in programs that combined housing and employment skills development. The policy implications of this “poverty trap” are universal, and have been highlighted in a Social Housing Services Corporation (SHSC) report in the Ontario context as an issue needing rectification.4 Given the successes of JobsPlus, the report suggested the need for more integration of housing and employment programs. • Location Contradictory evidence seems to point to differing effects of housing location on employment. Some authors highlight that public housing, located in low labour demand areas (i.e. high unemployment areas) can make it more difficult for a person to access work (Allard & Danziger, 2001; DTZ Consulting, 2006; Hulse & Randolph, 2004). While not looking specifically at public housing tenants, Dodson (2005) found a relationship between inequalities in labour markets and inequalities in housing markets that have exclusionary effects. For example, regions where jobs are available can often be accompanied by high house prices. These higher housing costs can be exclusionary; people who cannot afford the cost of housing may be forced to leave these higher-cost, job-rich areas for others with lower housing costs but where the employment prospects may be worse (Bridge, et al., 2003; Joint Center for Housing Studies, 2005; 4 Social Housing Services Corporation (2007). Snakes and Ladders: Ending Poverty Traps by Rebuilding Livelihoods in Social Housing. Toronto. 14 Mullins & Western, 2001a). Some of these low-cost regions are areas of public housing. This is somewhat echoed by Blunden, who points out that there are also findings which support the theory of spatial mismatching of employment, where willing workers do not find themselves located where jobs are readily available. Several American studies had findings that supported this conclusion (Ihlafeldt & Sjoquist, 1998; Allard & Danziger, 2003). Allard and Danziger found that social security recipients were more likely to work if they lived closer to work opportunities. However, the studies conducted on the MTO and HOPE VI programs to determine whether people who moved from areas of high poverty to areas of low poverty would be more likely to be employed found no difference between the experimental and control groups. This was also the case in other studies looking at the impact of living in areas of high poverty and employment (Buck, 2001; DTZ Consulting, 2006; Engeland & Lewis, 2006; Erebus, 2005; Levy & Wooley, 2007; Oreopoulos, 2003; Mullins & Western, 2001a; Phibbs & Young, 2001; U.S. Department of Housing and Urban Development, 2003). These mixed findings highlight the inter-related nature of all the elements, as there may be a locational factor involved in housing. But the research also points to such outcomes also being attributable to socio-economic status: i.e. simply not being able to move to “better” areas with more jobs as they might have unaffordable housing markets. According to Blunden, the Ihlafeldt and Sjoquist study also pointed to other factors such as mass transit accessibility and the concentration of poverty in a neighbourhood. • Stability In their 2004 study, Hulse and Randolph found that most unemployed tenants living in public housing felt that they had a better sense of security by being in public housing and could therefore have a better chance of looking for work. As with educational outcomes, the element of stability that public housing provides is likely an important factor for employment and skills development, a finding supported by Phibbs and Young. This sense of security reported by public housing tenants in all studies is important, most having reported an increased sense of self-esteem and economic stability that came from it. 2.3 Impact of Housing on Vulnerable Groups While the above observations and findings relate to the impact of housing on society as a whole, it is equally critical to try and understand the importance of housing to vulnerable groups often experiencing difficulty securing adequate, affordable and suitable accommodation. 15 “Links are being made from housing to the social exclusion of single parents, social assistance recipients, Aboriginal Canadians, new immigrants, and persons with disabilities. Housing is coming to be more widely seen as a key factor in social inclusion, connected to overall population health, healthy child development, and the creation of supportive and cohesive communities” (Jackson, 2004, p.7). The National Association of Friendship Centres (NAFC) (2004) identifies the high cost of shelter as the single biggest cause of poverty in urban Canada. According to a report by the Policy Research Initiative (PRI) Project (2005), roughly one in six Canadian households live in core housing need.5 In urban areas, both affordability and suitability play a role in the identified need; in areas adjacent to urban areas, affordability alone is the major concern; while in Northern rural areas suitable housing is identified as the greatest need. “Money spent on housing cannot be spent on other things – health services, recreation, education and nutrition” (Carter & Polevychok, p.12). The above-noted report also points out the role housing can play in facilitating and possibly magnifying the effectiveness of other supports for people living in poverty. To avoid marginalization, housing can provide individuals with the stability and physical security needed to move from a state of social exclusion to one of social inclusion with a new sense of order in their lives. The following discusses the findings from the literature regarding linkages between housing and education, skills development and employment for vulnerable groups. Although much research has been done on the housing needs of various vulnerable groups (such as seniors, youth and Aboriginals), it was found that little meaningful research has been undertaken on the impact of housing on the education-employment-skills development of specific vulnerable groups. This is, in fact, one of the biggest gaps that currently exists in housing research in Canada. 5 CMHC : A household is said to be in core housing need if it falls below at least one of the adequacy, affordability, or suitability standards and 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 three standards). Adequate dwellings are those reported by their residents as not requiring any major repairs; Affordable dwellings cost less than 30% of the total before-tax household income; Suitable dwellings have enough bedrooms for the size and make-up of resident households, according to National Occupancy Standard (NOS) requirements. A household is not in core housing need if its housing meets all of these standards, or, if its housing does not meet one or more of these standards, but it has sufficient income to obtain alternative local housing that is acceptable (meets all three standards). 16 • Aboriginal Peoples Much has been written on the housing needs of the Aboriginal population in Canada. According to Statistics Canada (2001), the highest concentrations of the Aboriginal population at the time of the Census lived in the North and on the Prairies. Eighty-five percent of the total population of Nunavut, 51% of the total population of the Northwest Territories, and 23% of the population of the Yukon are Aboriginal people. However, an increasing proportion of Aboriginal people are living in urban centres including Winnipeg (8% of the population), Saskatoon (9% of the population), Toronto (0.4%) and Montreal (0.3%). This trend has continued - a recent study by Statistics Canada (2006) on Aboriginal Peoples in Canada found that these numbers increased for the majority of cities in the 2006 census.6 Compared to the non-Aboriginal population, Aboriginal people living in urban municipalities are considerably worse off socially and economically in terms of their rates of homelessness, unemployment, poverty, and crime. They also tend to have a lower level of education, more health related problems, and have a larger proportion of single-mother led families among their population. It is also important to note that Aboriginal households were 1.6 times more likely than nonAboriginal households to live in core housing need (PRI 2005; 2001 Census data). Almost one in four off-reserve Aboriginal households were in core housing need, as are an extremely high proportion of Aboriginal single-parent households (Carter, Polevychok, 2004). Because of jurisdictional issues related to the access to services for Aboriginal people, it can be difficult for them to obtain appropriate services in urban centres (Graham & Peters, 2002). In their literature review, Carter and Polevychok (2004) found that housing remains a major problem for Aboriginal people who move to urban areas to find jobs and better housing. Several factors impede their ability to find suitable housing, which includes “the shortage of housing, discrimination by landlords, and lack of information on housing availability” (p. 7). As a result, they often end up living in declining inner city neighbourhoods, housed in poor quality and unaffordable units. Research showed that Aboriginal people were also over-represented among the homeless populations in every major city where statistics were available (Carter and Polevychok, 2004). They attribute the existence of such homelessness among the Aboriginal population to the “lack of appropriate policies, uncoordinated services, the absence of affordable housing, insufficient supportive housing, high unemployment, and cuts to welfare rates” (p. 10). This is compounded by other contributing social factors, including “poor health, mental illness, substance abuse, domestic violence and poverty” (p. 10). Several other issues were also identified which hinder access to services, such as a lack of transportation, childcare needs, lack of information, and lack of resources. 6 The 2006 Census found that Winnipeg, Toronto and Montreal’s aboriginal populations have grown to 10%, 0.5%, and 0.5% of the cities’ respective populations, with Saskatoon stable at 9%. 17 While the above studies and references speak to the wide range of housing concerns among the Aboriginal population, they do not provide specific research on the impacts of these housing problems on education, skills development or employment of this population. From our review, it is clear that, although there is a considerable amount of research regarding Aboriginal populations, very little exists in terms of research examining the links between housing and education, employment, and skills development for Aboriginal populations. This constitutes a major gap in Canadian housing research. • Women and Single-Parent Households The PRI Project (2005) cited a 2004 CMHC Report which points out that almost half of the non-Aboriginal single parent households with children under 18 living in rental accommodations lived in core housing need in 2001. In a literature review related to women and housing, Johnson and Ruddock (2004) found that women-led households have been over-represented in the proportion of lowincome families and those dealing with serious housing problems. The number of lone-parent households increased rapidly between 1996 and 2001, then increased more slowly between 2001 and 2006 (Statistics Canada, 1996, 2001, and 2006 Census). In Australia, Burke and Hulse (2002) indicated that “sole parents are arguably the most disadvantaged group” (p. vi) and are more likely than other types of households to live in poverty. Again, they found that these sole parent families are predominantly headed by women who were not in the work force because of childcare and support responsibilities. Women are also at a higher risk of suffering from family violence, making up roughly 85% or the victims of spousal violence (Chiodo et al., 2003). The severity of the violence women face is also worse than for men who suffer abuse or violence. For these situations, emergency and transitional housing is a key component, after which accessing stable permanent housing becomes important to re-establishing stability (SHS, 2005).7 As with other groups, the literature on victims of family violence and the housing outcomes is sparse, when relating to the specific education, skills development and employment outcomes that are the focus of this report. Women face particularly difficult housing circumstances, and in major Canadian cities there is an increasingly larger proportion of homeless women, including women with children. Households headed by women are more likely to be 7 This is somewhat echoed in a CMHC report from 2000 (Women on the Rough Edge) examining long-term homelessness in women, in which the lack of appropriate services for homeless women is highlighted. See also CMHC 2004, Transitional Housing: Objectives, Indicators of Success, and Outcomes. 18 renters and are disadvantaged in gaining access to home ownership, largely due to affordability.8 When women are looking for housing, safety is an important concern since, in some cases, women experience domestic violence in their homes. If women are able to find secure, affordable housing, their capacity to pursue education and professional goals and ability to form supportive networks is enhanced (Carter & Polevychok, 2004). This is supported by the research of Johnson and Ruddock, who found that the stability afforded by adequate housing was significant in aiding women to gain economic independence and help them enter the workforce. However, they mention the issue of childcare as also being an important element, one not considered within the context of our study. While the above research does show some specific linkages between housing and skills development and employment among women, the element of access to adequate supportive service for single-parent households is something to consider for future research. • Children and Youth The research in this area does point to some important conclusions about the linkage between housing and education, skills development and employment. The impact of low income families spending proportionately higher amounts of their incomes on housing is felt by the children of these families, since less money is available to send these children to summer camps, arts or sports programs or to engage them in other activities that children in higher income families have access to. This has long-term implications for the development of these children (Canadian Council on Social Development, 2002). In his discussion paper on housing affordability and children, Cooper states that “inadequate housing directly affects child health and well-being, and spending a large or disproportionate amount of income on housing means less money is available for other necessities.” (Cooper, 2001, p.17). In discussing the problems associated with finding affordable housing, Carter and Polevychok also point out that children can end up in the care of the Children’s Aid Society when families have problems paying rent, are evicted, have no permanent home, are transient, live in a shelter, live in an overcrowded setting or live in housing below the basic standard. This issue of quality of housing, independent of the neighbourhood in which the housing is placed, also impacts the psychological health of residents which, in turn, impacts non-housing outcomes such as employment and education (Evans, Wells & Moch, 2003). When children are homeless or living in slums or in noisy or overcrowded accommodations, there are negative education impacts. 8 Almey, 1996; CMHC, 2002, 1997; Johnson and Ruddock, 2000 19 Children living in stable housing have the opportunity to secure a better education in a stable environment (Carter & Polevychok, 2004). This is analogous to the findings on stability and housing quality and educational outcomes. • Seniors As with many other vulnerable groups, much has been written about the housing needs of seniors, but few observations were found about the impact of housing on education, skills development and employment. It may not be surprising there is little research on this since most seniors do not make these elements a priority at their stage of life as they are not usually seeking employment or education to sustain their finances. Over the next 30 years, seniors are expected to increase by 116% compared to 33% for the total population, meaning that by 2036 seniors will represent one of every four people in the population, compared to 2004 where they represented one of eight (Carter and Polevychok, 2004). In 30 years, Carter and Polevychok state that demand for seniors housing will increase by 117%, 140% for nursing homes and care facilities, and 115% for private dwellings (assuming constant 1996 tenure patterns). In 2001, according to Statistics Canada (2005) as cited in the PRI Project (2005), senior renters were identified as being among the groups most likely to be living in inadequate housing (approximately 43% in core housing need). This need is especially high in urban areas such as Toronto where over 50% of renting seniors were in core housing need. Hay (2005) points out that seniors have multiple housing challenges including “affordability, safety, isolation, maintenance, and so on” (p. 3). In May 2004, a task force report on seniors had eight recommendations that focused on providing housing and in-home supports to seniors. These include: income supports, program flexibility to allow seniors to earn employment income, expanding the federal role, and collaborative initiatives with community groups and the private sector (Government of Canada, 2004). While the above studies and references speak to the range of housing concerns among the seniors population, they do not provide specific research on the impacts of these problems on education, skills development or employment for this population. For future research it would be important to evaluate what, if any, relationships exist between seniors and the outcomes studied. Given that many projections are showing that the next generation will be living longer, healthier lives, and are already working longer into their “retirement years” than the previous generation, such considerations could prove interesting. From the literature review, it is clear that although there is a fair amount of research regarding seniors populations, very little exists in terms of research examining the links between housing, and education, employment and skills development for senior populations. 20 • Recent Immigrants and Refugees As with the other vulnerable groups which were identified, much research has been undertaken examining the issue of housing for immigrants and refugees. However, little research exists which specifically examines its impact on education, skills development and employment. In 2001, a Longitudinal Study of Immigrants to Canada was undertaken. It was the first such longitudinal study since 1970. Research using this information is currently underway to identify what is known about immigration and housing. The government groups immigrants that arrive in Canada into the following categories: Economic Class, Family Class, Protected Persons or Refugees. Toronto, Montreal and Vancouver are the three urban centres where almost three-fourths of newcomers live (Wayland, 2007). According to the 2001 Census, today’s immigrants are more likely to live in poverty and depend on social services than were their predecessors and these poor outcomes are reflected in their housing situations. It also highlighted that 36% of recent immigrants were living in core housing need, compared to 13.7% for non-immigrant households. Those living in large metropolitan areas, renters, and the most recent immigrants are those with the greatest core housing need. Although most immigrants find housing quickly and 18% of them live in owneroccupied housing only six months after arrival (Mendez, Hiebert, Wyly, 2006), there is wide variation in housing experiences correlated with immigration class, country of origin and other variables (Wayland). According to Wayland’s research, the biggest housing-related barrier immigrants face is affordability, which is exacerbated by the declining availability of non-market or assisted housing. For new immigrants, finding adequate housing plays a key role in their ability to integrate socially and to gain adequate employment. New immigrants settle mainly in larger, high-cost, urban areas like Toronto and Vancouver making housing affordability a major problem: “housing is an important reason why many immigrants are poor” (Jackson, 2004, p. 53). The core housing need of recent immigrant households is, on average, 4.7% higher than non-immigrant households (PRI, 2005). Jackson (2004) cites research by CMHC (2003) which identifies the fact that major issues for new immigrants are the scale of rents and the affordability and adequacy of housing in big cities in Canada. Hulchanski (1997) highlights discrimination problems reported by new immigrants when searching for rental housing. Immigrants’ access to adequate, affordable housing helps create the circumstances and/or opportunities to access other formal and informal supports and networks (Carter & Polevychok, 2004a). In an environmental scan of 21 immigrants and refugees, Douglas (2005) points out that immigrants and refugees, as a group, are, at many levels, among the most marginalized in Canadian society. She points out that the difficulty with finding affordable housing impacts other areas of settlement for immigrants, including childhood care, education and food security. In a review of neighbourhood effects, Oreopoulos findings echoed those of the previous authors. He found that many recent immigrants to Canada live in highpoverty neighbourhoods, starting out in poor immigrant enclaves, but then usually move on to more affluent neighbourhoods where their population share scarcely differs from that of the city as a whole” (2005, p. 8). Thus, Canadian high-poverty neighbourhoods are more often home to recent immigrants than their U.S. counterparts, with most of these residents moving out of the neighbourhoods within five years. Interestingly, Tremblay et al’s research found that neighbourhoods with a high proportion of recent immigrants had slightly higher educational performance than neighbourhoods that did not. These findings are all quite interesting, but suggest more work needs to be done on assessing the impact of housing on immigrants during their initial settlement period. Mendez, Hiebert and Wyly (2006) point out that, while more research is being undertaken to look at the varied housing experiences of immigrants entering Canada through different categories of admission, more complete research is needed. They also identify a problem with much of the research since the studies rely on different methodologies, making it impossible to undertake systematic, comparative analysis of the relationship between housing conditions and admission class. Most lacking, according to the researchers, is information about housing experiences of newcomers in their first six months in Canada. • Housing and Homeless Populations Housing and homelessness is a complex issue, especially as it encompasses those that are homeless, and those that are at risk of homelessness. Here again, while much has been written about the reasons for homelessness, the literature provides little specific research about the impact of housing on the education, skills development and employment of this population. The PRI Project (2005) defines homelessness as “a state of instability and exclusion from physical capital that can be a product of persistent poverty and that accentuates the negative effects of that socio-economic situation” (p. 4). The authors also cite Begin et al. (1999) who define homelessness as a “state of being that is not a characteristic of an individual but is rather a life situation that may be temporary, periodic, or more or less permanent” (p. 4). Hulchanski (2005) states “what we call homelessness is not simply a housing problem, but it is always a housing problem” (2005a, p. 3). 22 The process of becoming homeless consists of many stages and affects a wide range of people, including families with children. The process is complex, nonlinear and unpredictable, often combining concurrent issues such as substance abuse, family or sexual violence, or mental illness. This combination of issues makes examining, and effectively dealing with the issues of the homeless population particularly challenging. Those who are at risk of becoming homeless are those families and individuals with formal shelter but in precarious circumstances. What plays a major role in the complex process of homelessness is housing affordability. According to Hulchanski (2005b), before the 1980s, homeless people living in Canada were largely men, irregularly employed, generally transient, detached from their family homes and living in flophouses, rooming houses and other low-cost, low-quality places. Homelessness increased in the 1980s and affected diverse groups of people with diverse problems who are unable to find suitable, affordable housing. The importance given to stable housing for homeless populations is evidenced by endeavours such as Housing First, which prioritizes finding sustained or stable housing for people first and foremost. After this, appropriate services are made available. Few empirical studies were found that track the importance of such programs, and initiatives like Housing First would greatly benefit from studies into the outcomes of housing for homeless populations. Hwang also indicates that homeless youth or street youth (the terms are used interchangeably) refer to teenagers and young people below the age of 25. Toronto, Montreal and Vancouver have the largest numbers of homeless youth in Canada and most of them do not sleep in homeless shelters. He also points out that Aboriginal people are over-represented in the homeless population of Canada by a factor of about 10, stressing that they also constitute a disproportionate number of homeless people who sleep on the street instead of shelters. Josephson’s 2004 report indicated that, after a review of print and electronic literature, little documented Canadian research existed that related housing and homelessness to education and employment. He found that concrete theoretical models relating education and employment to homelessness are rare in the Canadian literature. He pointed out the need to use more empirical research methods to direct future research and trial interventions in this domain, including longitudinal and multi-site research methods. As noted earlier, much of the existing research and literature relating to vulnerable groups point to linkages between housing and education, skills development, and employment. However, as shown above, there is little direct research on how housing impacts specific vulnerable groups with respect to societal outcomes of education, skills development and employment in Canada. 23 Sections Three and Four of this report will discuss methodologies through which these research gaps can be addressed. 24 Family income / and educational attainment • Motor and social development; neighbourhood affluence/ poverty; family income; parent(s) education level; neighbourhood family structure. Education Examples of Indicators Used Neighbourhood affluence • Neighbourhood affluence; home / educational attainment ownership rates; residential stability/instability; achievement scores; literacy of parents; median family income; availability of books in household. 25 • Overall, findings are somewhat inconclusive: many authors stress that individual and family characteristics cannot be completely discounted, while others highlight incomplete knowledge of these linkages. • Some studies have pointed to higher family educational attainment and income as associated with educational outcomes. (Edwards, 2005; Kohen & Hertzman, 1999; Peck, 2001) • Other studies point to little or no effect of neighbourhood affluence. (Currie & Yelowitz, 1999; Jacob, 2003; Phibbs & Young, 2005; Sanbonmatsu et al., 2004; U.S. HUD, 2003) Conclusion and Strength of Linkage • Some studies have found positive educational outcomes for students and children related to neighbourhood affluence and level of neighbourhood education. (Beauvais & Jenson 2003; Buck, 2001; Edwards, 2005; Ellen & Turner, 1997; Erebus, 2005; Jencks & Mayer, 1990; Kohen & Hertzman, 1999 and with Brooks-Gunn 1998; Manski, 1993; Pebley & Sastry, 2003; Tremblay et al. 2001) • Noted as having a significant positive effect on drop-out rates of female students. (Kling & Liebman, 2004; Popkin, Leventhal & Weismann, 2006) The following section summarizes the linkages that were identified in the literature, and, where possible, briefly discusses the strength of the relationships that were found. 2.4 Summary of Identified Linkages • Child education performance; condition of housing; tenure; family income; educational services; Examples of Indicators Used Housing Conditions (physical characteristics: overcrowding, density, physical condition of dwelling, noise levels, etc). Employment and Skills Development Tenure (owning versus renting) Employment and educational level. • Success at school; family income; labour force participation; tenure; condition of home; civic engagement. Tenure (owning versus renting) • • Family income; number of family moves; number of times children changed schools; tenure type; condition of housing. Stability and Transience 26 • Lack of evidence and research showing clear relationship between employment/skills development and tenure. • Tenure is more indicative of SES and employment • Improved educational outcomes arise from housing stability. (Braconi, 2001; Kohen, Hertzman & Brooks-Gunn, 1998; Cooper, 2001) • Number of moves can affect likelihood of graduation and/or educational performance (Braconi, 2001) • Positive association with children’s educational attainment and home ownership in several studies. (Braconi, 2001; Bridge et al., 2007; Crawford & Londerville, 2004; Curtis & Phipps, 2000; Mullins & Western, 2001a; Rossi & Webber, 1996; Steele & Sarker, 2005) • Some studies point to negative aspects, or lack of understanding of home ownership impacts for low-income households. (Blunden, 2005; Mendelson, 2006; Peck, 2001; Rohe, Van Zandt & McCarthy, 2001) • Strong evidence showing negative educational impacts from aspects such as homelessness, noisy homes, overcrowded dwellings, and living in poor housing conditions. (Braconi, 2001; Cooper, 2001, Mullins, Western & Broadbent, 2001; Mullins & Western, 2001a; Phibbs & Young, 2005) • Evidence that good housing conditions are important for children’s educational outcomes, and can have repercussive effects (positive or negative) in other areas such as health and employment. (Evans, Wells & Moch, 2003) Conclusion and Strength of Linkage • Welfare receipt; neighbourhood poverty rates; job accessibility as jobs per job-seeker. Spatial Inequalities Neighbourhood affluence • Gender of household lead; level of / level of education education; parental earnings; sibling income comparisons over time; labour market engagement • Education level; skills level; employment earnings. Income structures and rent 27 status than it is causative. (Blunden, 2005) • Research has found a relationship between income-related rent structures and disincentives to work. Referred to as a “poverty trap” by some, can penalize tenants for working, especially in lower-paying jobs. (Hulse & Randoph, 2004; Phibbs & Young, 2005; Riccio, 2007) • Review of Jobs-Plus program in the United States highlighted positive outcomes for participants. Especially strong outcomes for women who participated in such endeavours. (Bloom, Riccio & Verma, 2005; Johnson & Ruddock, 2000; Kramer, 2000) • Housing location can have exclusionary effects on employment: areas of low-employment usually mean high-affordability, and areas of highemployment often have low-affordability. (Allard & Danziger, 2003; Dodson, 2005; DTZ Consulting, 2006; Hulse & Randolph, 2004) • Studies have highlighted that people are more apt to work if situated closer to work opportunities. (Allard & Danziger, 2003; Blunden, 2005; Ihlafeldt & Sjoquist, 1998) • A locational factor exists between housing and employment, but authors stress that other factors should also be studied in future work, such as transportation accessibility and concentration of poverty. • Studies have found few effects, either positive or negative, of high-poverty or low-poverty neighbourhood on labour or employment. (Buck, 2001; DTZ Consulting, 2006; Engeland & Lewis, 200; Erebus, 2005; Levy & Woolley, 2007; Oreopoulos, 2003; Mullins & Western, 2001a; Stability • Rental type (private or public); current employment status; living arrangement; education level; difficulty in paying rent; factors affecting decisions about employment; difficulties in finding employment. 28 Phibbs & Young, 2001; U.S. HUD, 2003). • Studies have found positive effects of stable housing in unemployed tenants, inferring better chances or outcomes in finding employment. (Hulse & Randolph, 2004; Phibbs & Young, 2001) • Some evidence points to stability having positive effects for women, facilitating support networks and aiding in accessing the workforce. • (Johnson & Ruddock, 2000) • Too little research on this issue at the moment to conclusively relate stability and employment. 3. Methodological Approaches Most of the studies in which the authors conducted literature reviews commented that the findings to date are generally not based on rigorous research and analysis. This was also highlighted in the study on The Evolution of Ideas about the Relationship between Housing and Economic Development (Godwin, 2005), which noted a problem with “the way in which the economic impact of housing has been viewed, rather than the exact nature of that impact.” The following Section addresses the methodological issues of the reviewed literature from Phase One. Comments pertain primarily to the strengths and weaknesses of methodologies, including the approach, the data and the reliability of conclusions. A more complete discussion of the various methodologies can be found in Appendix 1. 3.1 Methodologies and Data A review of the research literature finds a broad range of methodologies and analytical techniques utilized to better understand the linkages between housing and education, skills development and employment outcomes. The strengths and weaknesses of various approaches are assessed below. 3.1.1 Methodologies Reviewed Most methodologies used in examining the linkage between housing and education, employment and skills development revolve around informal approaches such as literature reviews, opinion panels, and surveys. This presents a particular challenge insofar as few of the studies referenced in the literature incorporate in-depth statistical and analytical techniques, which can cast doubts over the strength of the linkages highlighted in such studies. This observation emphasizes the need for statistically-based analysis when examining the existence or absence of such linkages. Multivariate analysis9 is an important, almost necessary tool, enabling one to control for differences in nonhousing characteristics of the households. The methods by which this can be achieved differ. The following briefly lists the range and key strengths and weaknesses of the methodologies used in the studies reviewed during the first phase of the present research: • Non-Experimental Studies 9 Multivariate analysis: “an approach widely used in marketing research due to the complexity of most marketing problems, where several factors are operating together, when one wishes to estimate the influence of each of the variables on the end result, e.g. in monitoring a test market, devising a media schedule etc. Discriminant analysis and Factor Analysis are the two best known and widely used multivariate techniques” (Baker, 2002). 29 • – – – Cross-sectional studies Examines a single moment in time Weakest for determining causation Many limitations • – Case-control studies Examines cases exhibiting the studied behaviour, and identifying similar groups (controls) who do not Stronger than cross-sectional studies for demonstrating causation but vulnerable to error Valid results are dependent on the appropriate choice of the control group – – • – – – Longitudinal studies Cohort is identified and a variety of factors and baseline characteristics are determined Cohort is followed over time and the study population is monitored Superior to either cross sectional or case-control studies in determining causation The majority of the studies reviewed were of non-experimental nature, many being mixed method studies which used quantitative and qualitative data, typically survey data. Another common methodology was the longitudinal study. These types, when appropriately controlling for weaknesses or bias, were generally methodologically sound. Many of the literature reviews involved a review of such studies as part of the analysis, • Experimental Studies • – Randomized control studies Subjects are randomly assigned to receive or not receive the treatment of interest – Subjects are followed over time – Best at determining evidence for a cause and effect relationship – Practical and ethical problems when applied to housing and education, skills development and employment – Expensive and time-consuming to conduct • “Natural” randomized studies – Subjects “naturally” assigned to treatment group or not due to administrative or policy-related decisions – Promising, as they can generate data akin to that of a true experimental study – Can generate valid results – Care must be exercised in ensuring that the groups truly are distinguishable 30 The MTO and the HOPE IV research studies were the only ones reviewed based on a true randomized control methodology. However, the work done by Oreopoulos on public housing residents in Toronto is a good example of a natural random assignment study. For an overview of the studies deemed most methodologically sound and their findings, please see Appendix 3. The findings of these studies are that there is little, or only a slight, linkage between neighbourhood change and the outcomes of interest. It seems plausible that other studies might show a stronger linkage, but these studies set a rigorous baseline for testing of such linkages, both in the collection of data and in the analysis of the results. They also set a baseline for testing of other linkages. 3.1.2 Available Methodologies and Data Sets: Challenges of Attributing Outcomes to Housing A more complete overview and discussion of the various available methodologies was compiled by Phibbs and Young (2002), who identified the advantages and disadvantages of five different empirical research designs that were used to study the relationship between housing and non-housing outcomes. Their description of these is contained in the table below: Study Design Advantages Comparative study of different communities Results are more immediate. Relatively inexpensive. Comparative study of different groups e.g. between housed and homeless populations Comparative study of rehoused/nonrehoused groups10 Research Limitations Difficult to match the socio-economic and demographic characteristics of the two groups Does not model the process of providing housing assistance Sharp focus between Matching the two groups non-shelter benefits of may be difficult two groups. Difficulty with isolating Relatively short study. the impact of housing The two resident samples Improvements in housing are likely to be more quality may be socio-economically confounded by location homogenous (although effects. there might still be some Only examines one type bias issues). of housing tenure (social Relatively short study. housing). May be difficult to find a large sample. 10 Where one group maintains the original housing, the other group is/has moved to higher quality housing in another area. 31 Provide excellent insight into the mechanisms that generate costs and the linkages between costs. Can reveal the trigger mechanism for the additional costs of unmet housing needs & act as an educative tool for the community and decision makers. Mirrors the activity of the Longitudinal Study provision of housing. Provides the most compelling evidence especially for those not familiar with social science research methods. Sharp changes in housing circumstances provide the opportunity to generate measurable changes in non-shelter benefits. Can examine the nonshelter benefits of a number of tenures (e.g. private rental, public housing, community housing) Source: (Phibbs, Kennedy and Tippett, 1999) Case Studies Difficult to extrapolate community-wide or whole of governmental costs from individual case studies. Decision-makers looking for traditional surveybased evidence may resist findings from such a study. Relatively expensive. If a prospective method is used the study will need to be reasonably long. Only examines nonshelter benefits of those who have been provided with housing assistance. Given this overview, the majority of studies reviewed which were deemed to be methodologically sound (whether links were proven to be strong or not) were empirically-based or were of a mixed method type. Data and Data Sets There are two types of data, experimental data and non-experimental data. The “gold standard” (Steele, 2008) is experimental data. • Experimental Data Experimental Data can be separated into two sub-types: 32 True Experimental∗ An example would be an experiment where half a group of renters is randomly selected to be assigned a home they own, while paying the same monthly amount they did as renters. The other half would stay as renters. The first is the treatment group, the second is the control group. Outcomes such as employment would be measured at different points of time for both groups. The MTO study is a well known example of such data collection and methodology. The main problems with such experiments are their expense and the complexity that can arise from controlling for many variables over time. Using the above example, the control group might decide to purchase homes over time, resulting in its attrition. It has also been stressed by Marion Steele that such experimental data in the social sciences may be far from the quality of data obtained in double-blind experiments as used by science and medicine. Evidently the nature of each realm of study does not lend itself to direct comparison, but rather because of these differences the quality of the data and the efficacy of using such experimental methods must be carefully considered. Natural Experiment∗ This type of data would reflect a situation where, for example, for a policy or administration reason one group is assigned a particular treatment while another is not. Such natural experiments can hold promise of generating data akin to true experimental, allowing for similarly powerful results. The study done by Oreopoulos used data from a natural experiment, in which public housing applicants in Toronto were randomly offered units in various kinds of developments. Such natural experiments are more apt to be used for the study of socialscience based issues such as housing, although they have their drawbacks as well. This is true for the problem of self-selection (see definition on following page), which can also cause similar problems of attrition in a group. For example, if social housing applicants were permitted to turn down offered units, those accepting co-op units might be those with more initiative. The two large randomized assignment studies, MTO and HOPE VI, are methodologically sound, even though the findings were minimal regarding the link between housing and educational outcomes. While the findings may not have shed great light on anticipated linkages, such rigorous baselines are necessary in setting a standard for the collection of data and analysis of the results. ∗ Steele, 2008. 33 • Non-experimental data Almost all data used in economics and other social sciences are nonexperimental, often produced using randomized surveys carried out by agencies such as Statistics Canada. The use of non-experimental data distinguishes the social sciences from the natural sciences. There are a number of important problems that one must bear in mind – whether the data are gathered in a randomized survey or not - when modeling and estimating with such data. These problems may exist whether cross-section or longitudinal data are used. They are the following11: • Endogeneity, or the problem that the independent variable may itself be affected by the dependent variable. For example, while whether or not a person is a homeowner may positively affect the percentage of weeks in the past year that the person was employed, this in turn may affect whether or not the person is a homeowner: a person who has spent many weeks this year unemployed may not be able to keep up his or her mortgage payments and may switch to renting. • Self-selection bias. For example, the people who become homeowners through the Habitat for Humanity program may be those low income people with extraordinary initiative, ambition, and a strong work ethic. These may be characteristics of desirable employees and those who will perform well in school. Thus, when this program is found to have a positive effect on educational outcomes, it may not be so much that the program has a positive effect as that there is huge selection bias: the people who select to participate may be the same kind of people who are likely to do well in school. Sometimes the self-selection problem can alternatively be called the unobserved heterogeneity problem. This would be the case where there is a sample of many individuals, some homeowners and some not, and the aim is to determine the effect of home ownership on the education outcomes of children. Some of the effect of home ownership in this case may be the result of the fact that people who select to be homeowners tend to be relatively competent and ready to plan ahead and these are important characteristics of parents who have a positive effect on the education of their children. This can also be called an unobserved heterogeneity problem where the heterogeneity is the competence and readiness to plan ahead of some, but not all, members of the sample. A problem is caused because this characteristic is correlated with the probability of being a homeowner. • 11 Correlation or the problem that variables included in the model may be correlated with variables excluded so that variables included in the model Steele, 2008. 34 • are picking up the effect of the excluded variables as well as their own effects. For example, if age is not included in the model as well as home ownership, then, since the probability of homeownership increases with age, the home ownership variable may have a more positive effect on the number of weeks worked than is attributable to home ownership per se. This would happen if weeks worked increased with age—which is probably true in a sample of people age 25 to 50. However, the problems of using non-experimental data have attracted a lot of attention from statisticians in the social sciences over the last few decades. First, to deal with the correlation problem, it is now well appreciated that many variables should be included in a model, i.e. that the model should be multivariate. The other two problems are in principle less tractable, but econometrics, which is statistics as applied to economic data, has as its major concern, solving such problems. For example, Heckman has developed methods to deal with biases in estimated effects caused by self selection (Steele, 2008). Econometrics has also developed methods to deal with endogeneity, a major one being the use of instrumental variables12. Such developments in econometrics and other social science statistical disciplines have meant that nonexperimental data can often be used to allow quite firm conclusions to be made. 3.2 What Data and Literature Gaps Exist? A key problem, according to Pomeroy (2004), has been that the focus of programs addressing housing needs in Canada has often looked at shelter outcomes, “particularly quantitative measures of units produced” (p. 2). He believes that a better assessment of how well housing investments are working would be to monitor the impact of housing interventions on non-shelter outcomes. This confirms the findings from Bridge et al (2003) in Australia, that there is a lack of empirical evidence to be able to examine and measure the possible impacts. “Without good data, ongoing research and a system of assessment, it is difficult to clearly identify which programs and policies have the greatest impact and payback for public investment” (p. 12). The challenge that the lack of reliable data represents is clear, as it can compromise the quality of the conclusions drawn from the results. In a 2007 paper, Riccio posits that, while a number of housing policy reforms have been tried to promote residents’ self-sufficiency, little of the innovation in the field (proposed or actual) is based on “credible evidence of “what works” making it difficult to know “whether self-sufficiency strategies for assisted housing that sound promising are really a good bet or a bad investment” (p. 1). Studies have been done to suggest that these value-added innovations are effective, yet these remain mainly descriptive, non-experimental studies that have data and/or 12 Instrumental variables: If a right hand side (RHS) variable is endogenous, an instrumental variable may be used to overcome econometric problems. (Steele, 2008) 35 methodological limitations, making it difficult to draw conclusions about their impact. Riccio argues the need for a stronger base of evidence through the use of randomized controlled trials in the housing-employment policy arena in order to enable the development of evidence-based policy. In their paper Building Capacity: Enhancing Women’s Economic Participation Through Housing, Johnson and Ruddock (2004) point out that there are few resources that specifically connect the importance of housing to the employment of women, despite the fact that there are vast amounts of literature about women and housing, and women and employment. There is wide acceptance that the connection needs to be made, but there exists little research available demonstrating the link. Neighbourhood effects have different impacts on individuals at different stages in their lives (Ellen and Turner, 1997).13 While several ethnographic studies14 have been done to look at neighbourhood influence on infants and preschool children, Ellen and Turner stress that few quantitative studies exist to show the extent to which early childhood development is influenced by neighbourhood context. They also identify the need for research that includes the impact of “teachers, classmates, coaches and the parents and families of classmates and friends” (p. 849) on elementary school children. They believe that the neighbourhood environment plays an increasingly important role during a child’s elementary school years. Since adolescents spend more time with their peers and less time with their families, this is a time when the neighbourhood environment is most compelling (Ellen and Turner, 1997). Again, many studies have been undertaken to explore what impact this may have on various elements of adolescent life; however, Ellen and Turner point out that “few studies in any of these areas effectively control for the endogeneity of neighbourhood location” (p. 853). Although this latter element is likely due to weaknesses in study design, it does impact the quality of the data collected. 13 One possible method of examining these issues would be to use a life course approach, see section 4.1.4 14 Ethnography is a form of research focusing on the sociology of meaning through close field observation of sociocultural phenomena. Typically, the ethnographer focuses on a community (not necessarily geographic, considering also work, leisure, and other communities), selecting informants who are known to have an overview of the activities of the community. Such informants are asked to identify other informants representative of the community, using chain sampling to obtain a saturation of informants in all empirical areas of investigation. Informants are interviewed multiple times, using information from previous informants to elicit clarification and deeper responses upon re-interview. This process is intended to reveal common cultural understandings related to the phenomena under study. These subjective but collective understandings on a subject (ex., stratification) are often interpreted to be more significant than objective data (ex., income differentials). (Garson, 2006) 36 In looking at studies related to the effects of neighbourhood on adults, Ellen and Turner indicate that most of these studies emphasize the importance of physical distance, rather than social composition, when looking at the relationship between neighbourhood residence and the labour market. They conclude that the impact of neighbourhood influences on adult outcomes may not be that great, based on their review of the literature. However, they go on to say that this conclusion may be based on the fact that “less work has been done exploring neighbourhood influences on adult outcomes” (p. 854). In addition, Ellen and Turner identify a gap in the empirical research literature studying the impact that living in a high-crime neighbourhood has on the emotional and moral development of young children, as well as on the risk of individuals becoming involved in criminal activity. Conclusions regarding tenure-types are mixed, and more studies examining tenure and societal outcomes would greatly benefit overall housing knowledge. Homeownership is not “necessarily an assured road to riches, or even to a moderately improved level of wealth, for all low-income families” (Mendelson, 2006, p.36). Rohe, Van Zandt, and McCarthy (2001) indicated that little is known about “the social-psychological or economic impacts of mortgage payment stress or mortgage default; that is, the role of home ownership in potentially trapping persons in neighbourhoods that they would rather leave; and the relationship between home ownership and efforts to exclude minorities, renters and others from neighbourhoods” (p. 25). These issues of tenure-type and outcomes are of significant interest as many of the findings do point to a correlation between places, such as neighbourhoods, and positive outcomes for those living there. However, as Mendelson commented on his findings, the outcomes may cause us to re-evaluate certain assumptions: in his case the vision of the house as an almost certainly good investment. Finally, as was discussed in Section Two, there are large knowledge gaps regarding the impacts of housing on vulnerable groups. The greatest lack of knowledge pertains to Aboriginals, seniors, homeless populations and recent immigrants and refugees. Some research has been carried out on women and single-parent households, as well as children and youth, but it is clear that more needs to be done. 37 4. Addressing Weaknesses and Filling the Gaps: Future Avenues As was discussed in the previous Section, much has been written, often by the authors themselves, as to how to address the issues of weaknesses in methodology and data collection or simply data availability. There is more to addressing these issues, as there are other important factors to consider, such as fostering research and ensuring adequate financial support, especially for larger studies such as the randomized assignment and longitudinal sample types. The following Section examines these issues, as well as discussing avenues to help future research on the linkages between housing and education, employment, and skills development. 4.1 Strengthening Methodologies and Data Collection Although many of the current methodologies are sound, there are certain issues regarding the choice of analytical models, as well as definitions that have been highlighted as areas for improvement. 4.1.1 Defining the Issues One of the challenges in examining variables such as the housing impacts on non-housing elements is the definition of exactly what constitutes such variables. In the case of neighbourhoods, it was previously stressed that the term itself encompasses a wide variety of elements which can vary greatly, and defining a neighbourhood is ultimately a subjective process (Hulchanksi, 2007). This issue of defining, or redefining neighbourhood boundaries was highlighted by several authors, including Ellen and Turner (1997), who used American census tracts, but indicated that this data may not accurately represent the neighbourhood conditions that impact people’s lives. In Canada, contrary to Ellen and Turner’s opinion for the U.S., Hulchanski’s conclusions in his recent study of income segregation in Toronto were that census tracts in Canada were generally a good proxy from which to study various neighbourhood level issues and outcomes. Statistics Canada’s use of the census tract as a neighbourhood proxy in its study on neighbourhood influences should also be noted. Many of the neighbourhood based studies rely on the census tracts as such proxies, and there appears to be a consensus on their reliability. Ellen and Turner’s empirical research into how neighbourhoods affect families and children in America, pointed out that there is little empirical evidence about 38 “the causal mechanisms through which neighbourhood environment influences individual outcomes” (p. 833). This indicates the need for future empirical research to take on the critical question of how and for whom neighbourhood matters in order to be useful to policy makers. They do note, however, that efforts to identify which neighbourhood characteristics matter most and to quantify their importance for their residents have been inconclusive overall. While educational attainment and employment may be influenced by neighbourhood conditions, most of the empirical research focuses on one of these outcomes at a time. They also supported the importance of using more than a “point-in-time measure” to study neighbourhood effects. The researchers point out that most of the empirical research they reviewed assumed that there is a linear relationship between neighbourhood characteristics and individual outcomes and nonlinearities are not tested for in many studies. They cite Crane (1991), who shows that neighbourhood effects operate like epidemics in how they spread, essentially saying that the prevalence of social problems should be much greater in areas that have experienced an increase in the incidence of the problem. Canadian studies have also pointed out the need to clearly differentiate between what are assumed to be reasonable outcomes, and what is done within the study itself to account for various elements (Statistics Canada, 2004; 2003; Ross et al., 2000). In many cases, the conclusions point to the importance of neighbourhoods for some outcomes, such as health (Dunn et al.,2006), but in others the authors cautioned that such linkages are not “automatic”, being mitigated by other factors (Ross et al., 2000, p.898). According to Sampson, Morenhoff and Gannon-Rowley, defining neighbourhoods based on census tracts or higher geographical aggregations is problematic when studying social processes. Since what happens in a neighbourhood is influenced not only by what happens in the immediate neighbourhood, but also by what happens in surrounding areas, their discussion included expanding neighbourhood research data collection to nearby areas outside the formal boundaries of a given neighbourhood. This has been echoed by some research at Statistics Canada, indicating that using census tracts may attenuate some associations, since “they may not coincide with residents’ perceptions of neighbourhoods” (2003, p.27). The caveat being that lower-level data, such as enumeration areas are not always possible to use. By focusing only on outcomes internal to a given neighbourhood, researchers are only getting part of the story of what influences non-housing outcomes. Sampson et al. suggest that these spatial dynamics are another way to measure child well-being, among other factors. The use of a “geography of street patterns” (referred to as tertiary communities15) is suggested when looking at 15 The authors are referring to work by Grannis that examined micro-boundaries within neighbourhoods. These “tertiary communities” are essentially aggregates of city blocks not bisected by major thoroughfares: this work highlighted, for example, how some residents 39 child well-being, since there are spatial constraints on children’s daily activity patterns, such as not crossing major roads in a neighbourhood. Neighbourhood indicators of child well-being and social processes could be developed based on these tertiary communities. Of particular concern to them is the use/misuse of control variables when studying neighbourhood effects which, they believe, may partition out “relevant variance in a host of mediating and developmental pathways of influence” (p. 469) since they provide a static model of measurement. According to the researchers, there are long-term community influences on non-housing outcomes that cannot be measured by controlling variables in the context of current neighbourhood characteristics of residence. “The general misuse of control variables in sociology thus appears to be exacerbated in the case of neighbourhood effects” (p. 469). They also warn against looking only at the characteristics of someone’s place of residence when trying to determine neighbourhood effects. They point out that many of the behaviours of interest take place outside of the neighbourhood in which these individuals reside. They suggest that “contextual theories that focus more on behavioural events than individual differences – for example, how neighbourhoods fare as units of guardianship or socialization over their own public spaces” (p. 469) should be utilized to measure neighbourhood effects. Ellen and Turner believe that individual and family characteristics need to be adequately controlled for in studies on neighbourhood effects in order to collect accurate data. While some individual and family characteristics are readily observable, such as income, education, and race, other relevant characteristics are harder to observe and are therefore not captured in empirical research. If these unobserved family characteristics are not adequately controlled for in research studies, empirical results may overstate the effects of neighbourhood. They cite the Moving to Opportunity program as “perhaps the ideal way to deal with unobserved family effects” (p. 847). This discussion highlights the fact that the current definition of what constitutes a “neighbourhood” is a difficult decision due to its subjective nature as Hulchanski stated. The current knowledge on housing outcomes, which does point to neighbourhood effects, also shows that “neighbourhoods” themselves may not be as big a factor as previously thought. Rather, it would seem that the various elements themselves come together, in some cases differently: Ross et al. highlighted such a difference between Canada and the United-States regarding income inequality and mortality. Their study showed that such effects in Canada are not as significant as in America, these differences indicating that other factors interacted more with people living within their tertiary community than others nearby, but were on the other side of a major thoroughfare, or other such obstacle. 40 such as the spatial distribution of social and economic resources may be of greater importance. Thus, the definition of neighbourhood may need to be revised to include other factors that were not attributed to it previously. Such factors include the individual characteristics mentioned by Ellen and Turner, as well as the wider geographical boundaries that Sampson, Morenhoff and Gannon-Rowley mentioned. This would of course bring its own set of challenges and complications, and would not be without its own difficulties, for example controlling for environmental and individual variables. However, an updated definition of neighbourhood would likely help in better understanding the complex series of relationships that exists within them, and that shape their communities. Such a vision will also help in re-evaluating certain assumptions regarding the effects of neighbourhoods. As research further develops on these issues, care and attention must be given to “how and for whom” neighbourhoods matter as Ellen and Turner stated. The complexities of what constitutes a neighbourhood and how the individuals fit into this context will help guide better, more accurate research. 4.1.2 Analytical Models The following section highlights some of the analytical models that appear to work better than others for measuring and examining the issues considered within the context of this paper. Experimental Models True experimental, randomized control studies are one of the best methods for collecting and analyzing reliable data. As was shown in the MTO and HOPE VI studies, they can conclusively show the presence or absence of a cause and effect type of relationship. However, their cost can often be prohibitive, and may not represent the best use of limited funding for housing research. In order to measure the effectiveness of work-focused interventions for individuals receiving various forms of housing assistance, Riccio (2007) suggests that randomized controlled trials provide a credible way to learn whether a social intervention is successful in producing its intended effects. However, Riccio states that random assignment is only a foundation for studying these effects. In addition, he cites the need for including a “comprehensive evaluation strategy with several strands of research, such as (1) an impact analysis, (2) an implementation and process analysis, and (3) a benefit-cost analysis” (p. 20). An effective impact analysis would use data collected from administrative records and surveys of residents and the impact would be measured by identifying the difference in an outcome between the program and control groups. Specified 41 subgroup analyses would result in being able to identify whether the effectiveness of an intervention is similar for different types of residents. Given the heterogeneity of assisted housing families and the dynamics of their lives, this type of analysis is important. An effective implementation and process study would be able to identify and assess how and how well the intervention model was implemented and the challenges associated with operationalizing the concept of the intervention. This type of study uses data drawn from qualitative field research and quantitative data received from agency records and surveys of residents. The results from both the impact and process analyses are incorporated into a cost-benefit study in order to estimate the resources used by the program and the economic outcomes it generates. Analysis of the return on investment includes looking at government budgets, taxpayers and the gains and losses to the program’s participants. Riccio warns that care must be taken to ensure that the interventions in random assignment studies are carried out as designed and that common conditions exist at the sites where the experiment is carried out. The random assignment process must also be implemented correctly in the field, which includes on-going monitoring. An adequate sample size is also necessary to determine if observed program effects are statistically significant and to be able to measure the impact of the program on a variety of different subgroups. Riccio identifies multisite studies as important for being able to assess whether the intervention can work in a variety of settings. Sobel (2006) confirms the importance of random assignment studies and supports Ricco’s contention that in a randomized study, the number of sites involved needs to be large enough to ensure that more precise causal parameters can be estimated. He also notes that it is important to collect data on social interactions before individuals are assigned to different groups in a randomized study to ensure that participants can be partitioned into equivalent classes with no direct social interactions, and treatments can be randomly assigned to these classes. The issue of interference and bias is also taken up by Sampson, Morenhoff and Gannon-Rowley in their 2002 meta-analysis16 of quantitative studies published in peer-reviewed or behavioural science journals. In their analysis, the authors identify selection bias as the biggest challenge facing neighbourhood-level research. While empirical studies such as Moving to Opportunity were designed to randomly assign families to one of three groups, there were differential take-up rates and dropouts from the program that impacted the results of these studies. They conclude that “while MTO may provide policy makers with evidence on whether offering housing vouchers can improve the lives of poor children, it is 16 “Meta-analysis is a statistical technique for amalgamating, summarizing, and reviewing previous quantitative research.” (Neill, 2006) 42 less satisfactory to social scientists interested in explaining the mechanisms of neighbourhood effects” (p. 467). Self-selection bias is identified by Rohe, Van Zandt and McCarthy (2001) in their study of the social benefits and costs of home ownership. They indicate that there is a significant threat to the validity of most of the research done on the impacts of home ownership because of the self-selection of people into home ownership and rental occupancy. They identify statistical techniques that can help account for the self-selection problem such as a two-stage modeling technique developed by Heckman (1979) that can be used to predict who becomes a homeowner based on social and economic characteristics. An independent variable is then developed based on the prediction to capture the effect of the selection bias in the primary regression model. They also point out that longitudinal research designs can address the self-selection problem since they allow the measurement of key variables before and after subjects become homeowners, enabling temporal sequences17, which are important in establishing causality. Haurin et al.’s 2002 paper on homeownership and child outcomes did employ a methodology that corrected for self-selection bias within their study by separately estimating the tenure choice of the parents, and the child outcomes which enables self-selection bias to be identified and corrected for. Mixed-Method Models In looking at causal mechanisms that link neighbourhood conditions to individual outcomes, Ellen and Turner (1997) suggest that qualitative methods “offer a promising approach for opening up the black box of neighbourhood effects” (p. 858). These qualitative methods include such approaches as in-depth, long-term participant observation in a neighbourhood, open-ended interviews with individual residents, and focus group interviews. While they point out that qualitative research can’t answer all questions about the possible influence of neighbourhoods on non-housing outcomes, it supplements quantitative research in critical ways by helping to disentangle neighbourhood effects from family and individual characteristics. “In conjunction with more generalizable results from quantitative analysis, qualitative research can both suggest hypotheses for further investigation and provide explanations for statistically significant results” (p. 859). One way to bridge the imbalance between the quantitative and qualitative approach to gathering data related to the interactions between housing and communities is through mixed-methods studies. Ellen and Turner (1997) identified such studies, combining both quantitative and qualitative measures, as a promising approach to the study of how and why neighbourhoods matter. Such an approach might help in furthering a conceptual framework for understanding how people’s behaviour and life chances are affected by neighbourhood 17 Temporal sequence: the logical order of causal event to occur (GMI, 2008) 43 characteristics: neighbourhood environments likely impact residents in different ways at different stages in their lives. Like Sampson, Morenhoff and Gannon-Rowley (2002), Ellen and Turner identify a number of contributing factors of neighbourhood life that might influence the outcomes of residents. These include: • • • • • • Quality of local services Socialization by adults Peer influences Social networks Exposure to crime and violence Physical distance and isolation A good example of this, on a micro scale, was highlighted in research done by Jeff May (2007) on the connectivity of Toronto social housing residents to various social service connections. Although the number of residents interviewed was small (18 in total), the results were nonetheless very interesting, highlighting most importantly the significant informal support networks that residents created for themselves, between themselves. Regarding employment and employment resources, the residents’ concerns had more to do with the low-wages of available jobs as opposed to the availability of the jobs themselves. Most of them made little use of the formal employment centres, mostly due to a lack of knowledge. The study could not account for larger neighbourhood effects within its scope, and could not examine empirically-based factors. However, it does beg the question, in this case and in the context of the present study, of what could have been a greater influence on employment: the neighbourhood effects, or those of the informal structures or other factors. Longitudinal Models Longitudinal studies are another excellent method, highlighted by several authors. Phibbs and Young point out that a longitudinal design is the preferred design to study the relationship between housing and non-housing outcomes. This approach should also incorporate a number of detailed case studies using a qualitative approach. Case studies assist in the framing of questions in the quantitative component of the study and in developing causal explanations. Sampson, Morenhoff and Gannon-Rowley also identify the need for “rigorous longitudinal studies of neighbourhood temporal dynamics” (p. 472). Because neighbourhoods change over time and are sometimes transformed, it is important for researchers to ensure that neighbourhood social processes are investigated in a “dynamic, interactive fashion”. 44 Other Models Buck (2001) has pointed out that there are a number of different models that researchers have used to identify causal influences of neighbourhoods on nonhousing outcomes. Outlined below is a brief description of these models: • Epidemic model – behaviour is thought to be contagious. Peer influences spread problem behaviour and different groups in the neighbourhood influence local social norms. • Collective socialization model – more successful adults provide positive role-models and are an important component of a child’s socialization. • Institutional model – the quality of services available in a neighbourhood impact on neighbourhood non-housing outcomes • Relative deprivation model – there is a reverse effect on non-housing outcomes when less successful people compare themselves to more successful people living in the same neighbourhood – these individuals fare worse in such situations. • Competition model – neighbours compete for scarce neighbourhood resources. • Network model – has an application to employment access. Effects arise from the presence of more affluent neighbours having a positive impact on neighbours who are less affluent. While the above models can be specified, discriminating them has practical difficulties. Buck points out the need to give careful attention to the selection of dependent and independent variables when researchers are studying the effects of a neighbourhood on non-housing outcomes. He cites a framework developed by Manski (1993) to identify the effects of social interactions which are separated into three types of influence: • • • Endogenous effects – the prevalence of certain behaviours in the group explains the propensity of an individual to behave the same way. Exogenous (contextual) effects – individual behaviour varies with the distribution of background characteristics in a group. Correlated effects – because individuals have similar institutional environments or have similar characteristics, they tend to behave similarly. Because people are influenced by their context at the same time as they influence the context, it is difficult, under normal circumstances, to separate these various influences in order to measure them. Another methodological issue identified by Buck includes a “standard omittedvariable problem” or correlation problem which occurs since “the existence of important unmeasured characteristics at the individual-level variables or arealevel variables may lead to biases in the estimates of area influences” (p. 2256). Buck also identifies the fact that people, in general, choose their neighbourhoods as another key methodological problem making it difficult to avoid bias based on 45 an individual’s unobservable characteristics that have influenced their choice of neighbourhood and outcomes. He also points out that the effects of neighbourhood can be mediated and moderated by actions taken by individuals or families, making it difficult to be able to identify true neighbourhood effects on non-housing outcomes. Because there are a number of different models to account for neighbourhood effects, Buck points out the need for care when including neighbourhood characteristics in any model. Buck identifies the following issues when studying neighbourhood effects: there may be non-linear associations between area characteristics and some outcome; the spatial scale at which neighbourhood effects are likely to operate must be considered; and using a uni-dimensional measure of area quality oversimplifies the analysis since there are a variety of possible causal pathways that lead to neighbourhood effects. Buck (2001) points out the limitations of non-experimental quantitative methods used to measure the impact of non-housing outcomes but also cautions that qualitative and experimental research also has limitations such as the possibility of participation effects and costs associated with experimental studies. He indicates that in the UK, better data at both individual and area levels, used to separate out level effects, and data properly related to plausible models of outcomes would provide effective ways to measure effects. He also suggests collecting data about migration and preferences. However, Buck reiterates the fact that, “individuals interact with their neighbourhoods in complex ways which may in the end make it difficult to disentangle the individual from the area either conceptually or in terms of data” (p. 2258). 4.1.3 Data and Data Sets The biggest problem hindering the research which was reviewed was the weaknesses and issues relating to data. In some cases, authors pointed out that this was due to a lack of recent, up-to date information. In other cases, researchers pointed out the use of incomplete data. Some of these issues arise from the relative expense of some research methodologies, such as the randomized control models like MTO and HOPE VI. In some studies, even methodologically sound work such as Phibbs and Young’s 2005 paper, Housing Assistance and Non-Shelter Outcomes, the issue was that they were only able to obtain a small sample size. In order to address such data gaps in Canada, Pomeroy (2004) suggests that there is a need for a more frequent collection of larger samples of data by Statistics Canada to be able to measure specific housing issues. Publication of this information should be produced in a timely manner. Pomeroy also identifies the need to broaden the indicators related to housing need measures to include non-housing outcomes. 46 Empirically-based research on non-shelter outcomes, according to Pomeroy, would enable more effective policy analysis and program design to help improve impacts and outcomes. By increasing the availability of data and identifying the need for information, Pomeroy believes academic interest and the development of research expertise would be generated. This research would provide added perspective and public debate about appropriate housing policies and ensure that there continues to be ongoing development of expertise of housing professionals. Longitudinal data sets A huge advantage of longitudinal data sets is the ability to follow persons and households through time. Two datasets include, for example, the National Longitudinal Survey of Children and Youth (NLSCY), used in the studies by Curtis and Phipps in 2001, and the Survey of Labour Income and Dynamics (SLID), one of many datasets used in Steele and Sarker. These datasets would be fertile ground for housing analysis, but some warnings are in order. First, both datasets, especially SLID, have a limited number of housing variables. However, Statistics Canada has added additional housing variables to SLID. Secondly, SLID and NLSCY survey a given sample of people for a limited number of years. Finally, the number of observations may not be large enough to mean that there are enough remaining once a sample is selected, for example, to focus on only those households who switched tenure in the first year or two of the sample periods (Marion Steele, 2008). The Longitudinal Administrative Database (LAD) The LAD is a dataset based on data taken from income tax returns, with additional variables added. It has a large number of observations. In provinces where there are property tax credits (e.g. Ontario) home ownership status would be available for a large number of observations. Many other variables are available on this database in addition to the income tax data. A major advantage of this database is that it is possible to follow a given individual for many years. Where housing outcomes of interest are long term, such as number of years contributing to the Canadian Pension Plan (an indicator of employment) the length of time individuals are followed in LAD is a distinct advantage. Cross-section data sets Cross-section public use microdata files (PUMFs) have been produced by Statistics Canada for over three decades. Among these are ones based on the Census and the Survey of Household Spending, and General Social Surveys (GSS). These include numerous housing variables. Their use to demonstrate cause and effect linkages on the outcomes of interest is limited by the fact they 47 refer mainly to a single point in time. However, it should be remembered that there are some retrospective variables in these data sets: for example, whether or not a household has moved in the last five years, number of weeks worked in the last year, and tenure of previous dwelling are all variables which have been included of these data sets. A major strength of these data sets is that there are a very large number of potential observations, as well as a large number of housing variables. Nonetheless, there are limitations on the usefulness of these kinds of data sets, and so Statistics Canada now produces longitudinal data sets. Another solution is to re-examine the use of cross-section data sets given their potential to provide instrumental variables to ameliorate endogeneity bias. Such instrumental variables are likely to have a small geographical dimension (e.g. a mean for a Census Tract) and their linkage to an individual would likely require the use of one of the secure Research Data Centres (RDCs) of Statistics Canada. It seems likely that the usefulness of these data sets for housing outcomes analysis is greater now than it used to be, because of the great range of data available through the RDC, and in any case these data have never been fully exploited. For example, although Steele and Sarker used GSS data, the scope of their project—considering a vast range of possible costs and benefits of home ownership—and a limited budget precluded doing full multivariate modeling of a few possible outcomes. Revisiting the data they used with more time spent developing models with additional independent variables and possibly using instrumental variables would be worthwhile. The Longitudinal Survey of Immigrants to Canada (LSIC) Mendez, Hiebert, and Wyly (2006) have undertaken to answer a number of research questions in their 2006 study using the Longitudinal Survey of Immigrants to Canada (LSIC). The authors point out that only the results of the first wave data are included in their study since second-wave data had not been released at the time of their study. Renaud, et al. (2006) also indicate that, with the release of second-wave data, additional research can be undertaken to look at the movement of immigrants in the “medium term” after they have spent more time in the country. They indicate that data from subsequent waves of the LSIC will enable researchers to “see whether the mechanisms observed are maintained, whether some are peculiar to the beginning of settlement, or whether other dynamics emerge after a certain time has passed” (p. 78). The limitations of the LSIC have been pointed out by Mendez, Hiebert, and Wyly. They indicate that, compared to the census, the LSIC has a small sample size which limits statistical study at fine levels of disaggregation and geographic scale. The LSIC includes only immigrants who were legally admitted and arrived in Canada during the survey period. Asylum claimants and refugees accepted through an asylum claim were excluded from the sampling frame of the survey. 48 They also discuss the potential under-reporting in the income and earnings questions in the survey and the problems with the methods used by Statistics Canada to predict the missing variable for non-responders. However, the LSIC does enable researchers to address some of the methodological obstacles faced earlier when studying housing and immigration in Canada. The Longitudinal Survey of Immigrants to Canada (LSIC) is the first survey of its kind for a generation (Mendez, Hiebert & Wyly, 2006, Renaud, et al, 2006). The LSIC provides a range of variables not included in the census which includes the housing search experience of immigrants, housing mobility, reasons for changing residence, newcomers’ socio-economic situation, motivations for immigration, labour market participation, integration barriers, access to health care and education, and settlement support sought and received from institutions and social networks. This data lends itself to future studies related to the housing situation of Canadian immigrants (Marion Steele, 2008). LAD was used by Oreopolous to determine the earnings outcomes of those who lived in Toronto public housing when they were teenagers. Methodological Notes of Caution The use of proxies, such as the poverty rate or average income levels, to study neighbourhood effects has been questioned. Ellen and Turner have indicated that such proxies may not accurately reflect those neighbourhood characteristics that are most important, and therefore would not enable policy makers to determine what interventions to target. They have also pointed out the extreme difficulty, if not impossibility, of differentiating different neighbourhood effects because of their high correlation. Another note of caution relates to the use of census data that “document neighbourhood-level correlations between percent recent immigrants and various measures of social problems (e.g., poverty, reliance on social assistance, low educational attainment, etc.) mistakenly infer individual-level relations from aggregate data” (Mendez, Hiebert and Wyly, 2006, p. 84). This highlights the point that assumptions based on this use of aggregate census tabulations mean that conclusions in this body of work are “tied to a variety of risky ecological assumptions” (Ibid., p. 84). In a 2006 paper, Sobel points out the importance of taking into account the possibility of interference when undertaking empirical research, such as the research carried out for the Moving to Opportunity (MTO) program. He identifies conclusions reached about these studies based on the “stable unit treatment value assumption (SUTVA)” which assumes that “the potential responses of units do not depend on the mechanism by which treatments are assigned or on the treatment received by another unit, i.e., there is no-interference between units” (p. 1398). Sobel identifies the fact that in studies with human subjects, social interaction is a primary source of interference. He describes the method by 49 which volunteers were recruited for the project, where many people who were recruited from group meetings would undoubtedly know one another. Sobel indicates the need to define effects when interference is present when undertaking empirical research. Finally, regarding the impacts of home ownership, Rohe, VanZandt and McCarthy (2001) have stressed the need for future research to do a better job controlling for alternative explanations for the relationships found. In addition, future research should strive to better identify the mechanisms through which home ownership influences various social variables. In addition, future research should strive to better identify the mechanisms through which home ownership influences various social variables. The researchers also found that much of the existing research on the impacts of home ownership does not recognize that the home ownership experience may be different for different types of home buyers, or for those who buy in different neighbourhoods or housing markets. They believe that there may be a bias, particularly on the part of American researchers, toward testing for evidence of purported positive impacts of home ownership. 4.1.4 Considerations for Indicators of Housing and Education, Skills Development and Employment Impacts It is also important in the assessment of how to address the weaknesses and gaps present in current research to review potential avenues for new indicators which may help in better designing studies, collecting data, and analyzing the results. This is partly due to the difficulties in measuring empirically for social or individual characteristics or behaviours that might be associated with neighbourhood effects. Another reason is the need to broaden the scope of what might impact the non-housing outcomes, which can change from one place and one group to another. The following elements could be used to further develop research and knowledge on the links between housing and education, skills development, and employment impacts. Neighbourhood Effects The difficulty in measuring “disparate but converging measures of neighbourhood mechanisms” (Sampson, Morenhoff and Gannon-Rowley, p. 457) that have independent validity, although related, is clear. Such factors would include social ties and interactions (i.e.the concept of social capital); norms and collective efficacy (i.e. the “linkage of mutual trust and the shared willingness to intervene for the public good”, p. 457); institutional resources (i.e. the “quality, quantity, and diversity of institutions in the community that address the needs of youth” p. 457); and routine activities (i.e. how children’s well-being is affected by land use patterns). As May’s (2007) research highlighted in the case of the interviewed Toronto Community Housing residents, many of these elements have not 50 necessarily been measured in the current research. This observation is also echoed in Josephson’s work (2004). Potential ways identified to measure such neighbourhood effects include the use of crime rates; the way that social ties are activated in a community (collective efficacy); and the measurement of social mechanisms and health including mental health outcomes and high-risk adolescent behaviours such as early sexual initiation, teen childbearing and conduct disorder (Sampson, Morenhoff and Gannon-Rowley). It has been mentioned that concentrated poverty and structural characteristics are still critical predictors of neighbourhood effects. Socio-economic resources and residential stability are necessary components of neighbourhoods that have high levels of collective efficacy. Maclennan (2001) has also identified the difficulty associated with trying to understand neighbourhood effects. Variables such as location, ethnicity, poverty, and neighbourhood deterioration are difficult to disaggregate when looking at the role that each plays. He points out that there are few statistical sources of information to provide suitable measures to be able to determine results and correlations. Statistics Canada also uses in its various data sets and products what could stand in for readily available indicators. For education, some potential indicators would include people having finished high school or not, and those with university degrees. Test scoring is another useful indicator. All of these already exist in some Statistics Canada data sets, along with housing information. Employment Related Elements Regarding skills development, years of schooling, trade certificates, changes in income, and changes in occupation groups to higher level ones can be obtained in micro data sets and the SLID. Finally, employment impacts could be evaluated by evaluating the number of weeks worked, increases in numbers of hours worked and changes in income. Social Indicators Based on the discussion in The Social Report (Ministry of Social Development, 2004), social indicators have been suggested as another avenue for measuring such outcomes. “Social indicators are signposts that help us to measure progress towards a desired outcome. Indicators are selected because they either directly measure the outcome of interest or because they are known to be a good predictor of, or are associated with, that outcome” (Ministry of Social Development, 2004). These indicators include measuring factors such as participation in early childhood education, participation in tertiary education, participation in sports and leisure, experience of cultural activities, telephone and internet access in the home, and regular contact with family and friends. 51 In the report The Use of Social Indicators as Evaluation Instruments, Ekos Research Associates (1998) identified a number of strengths and weaknesses of these methods. One approach, the use of social modeling through multivariate methods aimed at establishing causal relations, uses policy-relevant variables introduced as one independent variable, among others that contribute to the social outcome in question. The outcome is represented by variables that can be called social indicators. In studies they reviewed which examined child development, social indicators such as the level of social competence or schoolreadiness were used in conjunctions with socio-demographic and economic indicators. Ekos reported that advantages of this approach are that the models are generally based on some kind of conceptual or causal model where the links between inputs and the outputs – represented by social indicators – are clearly specified, as are the external control variables affecting these outcomes. The simulation model is based on a conceptual or theoretical model, ensuring that there is coherence and efficiency in the data collection and research. Explanatory variables include government expenditures making it possible to measure the contribution of programs to final social outcomes. These simulation models have the flexibility of posing and answering “what if” questions with rigour. One problem the researchers identify with social modeling approaches is that they are ordinarily confined to one area of social concern. Another challenge highlighted by Ekos is that social indicator efforts are purely quantitative: as a result, an individual’s valuation and judgement – how they feel about their social conditions - is not included in the measure. By including this qualitative information it would be possible to identify whether the observed changes are good or bad. They also point out the problem with the lack of an overall conceptual framework or theoretical model in the development of social indicators. Finally, an article by Armstrong, Francis, Bourne and Dussuyer (2002) stressed that another problem with indicators is that they tend to show the results but not the cost of the actions. Thus, the difficulty is complicated by the need to disentangle the causes of the change in an indicator. They identified the need for sophisticated modeling to disentangle the contributions of different programs, and it may still be that the causes may be external to the initiatives or the model. Social Capital Recent research on neighbourhoods has begun to point in the direction of social capital, the relationships that exist between individuals and their social networks, as an important source of neighbourhood change and stability. “Although social scientists are far from knowing if social capital is the answer to neighbourhood stability, the identification and assessment of the institutional and political, as well as socio-political, attributes of neighbourhoods may help to better explain the 52 dynamic process of neighbourhood change, condition, and character” (Nelson A. Rockefeller Institute, 1997). Pope (2003) explains that the use of social capital has become “one of the most popular exports from sociological theory into everyday language and has evolved into something of a ‘cure-all’ for the problems and challenges that confront societies and their modern development” (p. 1). Because the concept focuses on the positive aspects of human relationships, putting these positive consequences in the broader framework of capital and focusing on non-monetary capital as a source of power and influence, researchers find the concept persuasive. Pope also indicates that discussions of social capital also encompass the following elements: social connectedness, social cohesion, community competence, social networks, social inclusion, social support, social isolation, and social exclusion (p. 1). While social capital has an important role to play in explaining inequalities, according to Pope, she points to the fact that there is little evidence to date that social capital will provide a ready remedy for major social problems. In order to benefit researchers and policy makers, they will need to first examine the underlying theoretical basis of the definition of social capital they use and will also need to determine the types of measurements needed. Several criticisms related to social capital indicators were highlighted by Pope, the first being that social capital indicators lack clear definition, and that collective social capital is not the same as individual social capital. She argues that the presence of social capital may not always result in positive social outcomes, and solutions based on an individualized notion of social capital may not work or may reinforce inequality. Pope cites Labonte (1999) who warns that “solutions based on current social capital approaches of empowering people are as likely to fail as the many community development interventions that have gone before – if maximizing social capital is seen as the means to increase economic growth, and as a substitute for adequate infrastructure” (p. 9). Life Course Approach Finally, another option of interest is the life course approach, adapted from epidemiology and used in population health studies, which focuses on the trajectories of individuals through life (PRI, 2004). This approach examines elements cumulatively over a person’s life, but could prove to be an interesting method of examining neighbourhood effects of housing, accounting for the number of factors and complex relationships that exist.18 18 See Hertzman, C. and C. Power (2001). A Life Course Approach to Health and Human Development, in Jody Heyman, et al. (ed.), Healthier Societies: From Analysis to Action. Oxford University Press, pp. 83-106. 53 4.2 Future Research Considerations A range of suggestions have been made by various housing researchers to generate more useful research on the societal benefits of housing. These relate to issues of methodology, data, process, funding and partnering, as discussed below. Josephson (2004) emphasized that research needs to move beyond simple counts and descriptions to more in-depth explanations and the exploration of the relative importance of these factors. Given the current modest state of knowledge on the impact of housing on education, skills development and employment, several future research opportunities should be highlighted. A suggestion made by Pomeroy (2004) was that the need measurement developed by CMHC, designed to assess a core of specific housing issues, be undertaken more frequently and results published earlier. He also identifies the need to broaden the indicators of these data collection tools to include nonshelter outcomes. This would indeed help by making more data available to researchers. On this note, it could be suggested that Statistics Canada include more housing variables in its surveys. Several authors have highlighted the need for more recent empirical studies on a variety of issues. Any future research endeavours would be wise to enable or facilitate the development of such studies, especially when one considers the age of the data being used. It is recommended that further research be fostered to fund more specific, longer-term research initiatives aimed at collecting data for analysis on education, skills development and employment outcomes. These types of initiatives would benefit greatly from a national approach, examining similar issues across a range of cities and areas of the country. In an article that sets out, in part, to examine the validity of some of the research related to the sociological assumptions underlying housing mobility programs, Varady and Walker (2003), suggest that there is a need for research on the longterm impacts of helping families move to low-poverty, low-minority neighbourhoods, such as the MTO program. They identify the need to determine whether children in these programs experience future improvements in their employment circumstances due to better educational attainment. Other opportunities would include partnering with academic research groups. Such partnerships could be for longer-term research initiatives, as mentioned above. Such relationships would be of mutual benefit, enabling both sides to bridge funding issues, as well as helping in knowledge sharing: using existing information, programs and data as well as fostering new ones. 54 One example that merits to be highlighted is that such relationships could help in exploiting existing situations. A natural experiment of sorts is currently going on in Ontario at present and was used by Oreopoulos in his work: virtually all applicants for social housing are assigned to units on a first come, first served basis. It should thus, in principle, be possible to survey social housing occupants, as well as those of co-op housing and non-co-op housing at several points of time after their occupancy to determine, for example, whether changes in employment status relative to that at entry into the unit are better for co-op housing than other kinds of social housing. Finally, one of the most important gaps needing to be addressed is that of the lack of information pertaining to the non-housing outcomes of education, skills development and employment of the identified vulnerable groups. Of these groups, Aboriginals, homeless populations as well as recent immigrants are the groups in which the lack of knowledge is most evident. This, of course, does not mean that research pertaining to the outcomes for women and single parent households, as well as children and youth should not be a priority either. Adequate research of housing outcomes for these groups is of great importance given that they represent the groups at the margins of the housing system. The importance of, and difference that education, skills development and employment can make in lives of individuals, mean that understanding their links with housing is of great importance for these groups. One particular challenge, especially regarding the large, longer-breadth methodologies that were reviewed, is the lack of funding. Another challenge is the short-term nature of current funding which is problematic for researchers wanting to undertake evaluations of programs designed to support the needs of homeless people (Josephson, 2004). Josephson has identified the need for sustained funding in order to undertake detailed and objective research to design interventions and examine their effectiveness. Such calls are echoed by many specialists in the housing field, and are certainly an important obstacle to furthering housing research in Canada. The findings relating to neighbourhoods and their effects are appealing given the interest in creating mixed-income neighbourhoods. However, the differing conclusions regarding these links suggest that more research is needed to better understand the relationships that exist. Currently, the redevelopments of Regent Park and other properties of Toronto Community Housing into mixed-income communities could provide fascinating opportunities to better understand the complexities of neighbourhood effects. From a policy perspective, Smith and Torjman (2004) have stressed the “failure of horizontality” when dealing with complex files, in which there is a lack of intragovernmental collaboration on initiatives. In a discussion of the National Homelessness Initiative, it was reported that, while one of the essential elements 55 for approval of community plans was evaluating the outcomes of activities, little or no collaboration or guidance was undertaken on the methodologies for such evaluations. In addition, the money spent on these evaluations would be taken from the funding for the community services or projects to directly help homeless people. This also highlights the need for adequate funding dedicated to housing research, without compromising the effectiveness of current or future programs. Torjman points out that complex files require horizontal management. Measuring results from projects undertaken by different levels of government means the creation of new partnerships between and among organizations that had not worked together in the past. She also identifies the need to be able to measure both process and outcome when working within complex files. Different sources of quantitative and qualitative data need to be identified. In addition, longer-term outcome measures need to be established. 56 5. Concluding Discussion There is evidence that housing can and does have tangible outcomes regarding education, skills development and employment. However, it is also evident that there are many grey areas regarding these linkages. In many cases the outcomes were not found, were weak, or there was not enough evidence. In other cases methodological weaknesses, limitations regarding data or sample sizes affected the quality of the conclusion. With regards to educational outcomes, the strongest conclusions included neighbourhood outcomes, specifically the affluence of the neighbourhoods. Stability of housing was also found to have a strong linkage to education outcomes, and frequent moves or changing of schools was associated with poorer outcomes. Available parent time, the physical condition of housing, and housing stability were found to have an impact on education outcomes of children. Other conclusions, such as the outcome of tenure on education, the effects of homeownership, and living in public housing versus low-poverty areas, were less strong. The outcomes for employment and skills development that were the strongest included the effects of employment programs to mitigate income and rent structure disincentives. The effects of tenure were less conclusive having mixed results, however educational achievement and employment status did have links to tenure. Locational effects were also mixed, with some findings pointing to a relationship in labour market and housing market inequalities. However, the HOPE VI and MTO research found little evidence for this, and other findings attributed these locational inequalities to socio-economic status. These mixed findings highlight the difficulty of examining certain outcomes due to the complexity of factors involved. Finally, some evidence was found that pointed to stability of tenure of public housing tenants being a likely factor in employment outcomes, but could not be established with certainty. These findings are important in that they illustrate the need to more effectively study these outcomes and address research limitations that have been highlighted. A more complete, detailed and thorough picture of the full effects of housing will ultimately strengthen policies and programs. One of the challenges in examining the link between housing and education-skills development-employment is its complexity, as well as the impact of other intervening factors such as individual choice and community influence. These complexities cannot always be reliably quantified empirically and further research should consider multi-faceted approaches to aid in the understanding of these links. 57 Appendices 58 Appendix I - Proposals for Future Research Projects This Research Report has identified a number of gaps in research on the linkages between housing and education, skills development and employment. It has also identified the research approaches considered most suitable for providing fully supported findings and conclusions about these linkages. Based on these observations, the following outlines several research projects to increase understanding on the impacts of non-housing outcomes in Canada. 1- Non-Housing Outcomes in Ontario Affordable Housing Building on the work that was done by Oreopoulos, it was highlighted that a seminatural selection situation currently exists in Ontario: applicants for social housing are assigned units on a first-come, first-served basis. Such a situation could, in principle, be used to survey social housing occupants over a period of time after their occupancy, to determine if changes in employment, for example, were better for certain types of social housing units than others. If properly designed, this situation could be used to measure several elements of housing and the education, skills development and employment series of links. An addition to this research might be that of including a test for house condition effects. Another interesting extension or variation to this project would be to see whether similar situations in other provinces could be used to the same effect, broadening the scope to a more national or provincial level. To ensure a strong collection of information, it would be important that such a study be given a proper length of time within which to follow the individuals (perhaps a two to three year minimum). This would allow for adequate tracking of certain variables, especially those related to education and children for example. 2- Continuation and Follow-up of Longitudinal Sample Studies There are two major data sets which are of use, the National Longitudinal Survey of Children and Youth (NLSCY) and the Longitudinal Survey of Immigrants to Canada (LSIC). In the case of the NLSCY, research has found that ownership and housing condition did affect children’s education (Kohen, Hertzman & Brooks-Gunn, 1998; Curtis & Phipps, 2000). Such findings are relatively strong, though there are self-selection biases which must be controlled for, but nevertheless the issues that were found should be further explored. 59 Regarding the LSIC, several authors have highlighted its strengths, and with the third wave of data having been released more work will be done comparing the findings from the use of earlier data (e.g. Mendez, Hiebert & Wyly, 2006, Renaud, et al, 2006). As recent immigrants and refugees were a group for which information and research was lacking, such research work would be very beneficial. 3- RRAP Outcomes Study The Residential Rehabilitation Assistance Program (RRAP) could also be used for a shorter term study, in which households would be followed for various factors. This could include health and educational outcomes in children for renovated and deteriorated housing. One method would be to randomly approve acceptable applications for markets which are seriously oversubscribed, followed for a year for example. Another option could be to identify households that have been on waiting lists for several years and others that did not wait, comparing outcomes of variables. The use of the RRAP for such data collection could prove a good bridging measure until other programs or samples were amassed. 4- Education, Skills Development and Employment Outcomes in Aboriginal Populations Recent Statistics Canada data has shown that Canada’s urban Aboriginal population is growing, representing many challenges for all groups and governments involved. Given the great lack of knowledge and information regarding the non-housing outcomes for Canada’s Aboriginal populations, an important contribution would be a major study examining the educational, skills development and employment outcomes of housing for urban Aboriginal populations. Using the existing First Nations Regional Longitudinal Health Survey (FNRLHS) and Aboriginal Peoples Survey could prove to be important assets in better addressing the current lack of knowledge regarding the housing outcomes for Canada’s Aboriginal population. 60 Synthesizes findings from empirical research. Discussion paper. Includes Environmental scan. Beauvais, Caroline and Jenson, Jane (2003). The Wellbeing of Children: Are There “Neighborhood Effects”? METHODOLOGY Ellen, Ingrid and Turner, Margery (1997). Does Neighbourhood Matter? Assessing the Recent Evidence STUDY PROVEN LINKAGES & CONCLUSIONS • 61 Findings from the literature indicate that: neighbourhood effects are shaped by children’s different experiences by gender, class and ethnicity; social composition across neighbourhoods generates as much heterogeneity as is found among individuals; Determining causality is the most difficult challenge for researchers in this field. There is contradictory evidence as to which neighbourhood characteristics matter most with respect to educational attainment and which adolescents may be most affected by neighbourhood effects, although there is general support for the notion that neighbourhood does play a role in educational attainment. Strengths: Includes an environmental scan of interventions designed to improve child outcomes at the neighbourhood level. HOW WEAKNESSES HAVE BEEN ADDRESSED • WEAKNESSES/ LIMITATIONS OF APPROACH Difficult to identify and measure the neighbourhood conditions that play the most important role in shaping outcomes; neighbourhood effects may be non-linear, therefore not easily discernible; difficult to separate individual or family characteristics from neighbourhood qualities. Description: Updates the state of knowledge related to the impact of neighbourhood on child development. Strengths: Looks across outcomes to assess the overall importance of neighbourhood environment on the social and economic well-being of individuals at different lifestages. Description: Examines how neighbourhoods affect families and children. Identifies methodological challenges, summarizes past research. STUDY DESCRIPTION & STRENGTHS OF APPROACH Appendix II – Review of Identified Methodologically Sound Studies Curtis, L. and S. Phipps (2000) Economic Resources and Children’s Health and Success at School: An Analysis Using the NLSCY. HRDC. STUDY Longitudinal study, using regression analysis METHODOLOGY Strengths: Regression analysis using longitudinal data, only children present in both years of the survey are Description: Uses data from the second cycle of the National Longitudinal Survey of Children and Youth (NLSCY) to examine the relationship between economic resources and a child’s development. STUDY DESCRIPTION & STRENGTHS OF APPROACH Difficult to determine direction of causality. WEAKNESSES/ LIMITATIONS OF APPROACH Regress outcome measures using the first wave of the NLSCY to provide a benchmark; controlled for a number of variables using regressions; used twoperiod income date to find average poverty HOW WEAKNESSES HAVE BEEN ADDRESSED Children who live in owner-occupied housing have better outcomes than those who do not. More hours of parental time available each week significantly improves a child’s success at school • • 62 Neighbourhood and community circumstances are only one of three factors that affect child outcomes – the other two are income and parenting. Families and children living in supportive communities do better than those in communities lacking cohesion, good services and facing threats to well-being. • for some children, poor child developmental outcomes occur regardless of SES, neighbourhood or other circumstances. PROVEN LINKAGES & CONCLUSIONS Longitudinal study Longitudinal study. Analysis of Census data Buck, Nick (2001). Identifying Neighborhood Effects on Social Exclusion Engeland, John, and Roger Lewis (2004). Exclusion from METHODOLOGY Bloom, H.S., Riccio, J.A., Verma, N. (2005). Promoting work in public housing: The effectiveness of Jobs-Plus. MDRC. STUDY Description: Overview of issues pertaining to core housing need and its Strengths: Individual records are linked to census data. Description: Uses data from the British Household Panel Study linked to local area characteristics to examine how/whether non-housing outcomes are associated with neighbourhood characteristics. Strengths: Two resident surveys were conducted over time. Used random assignment research designs. included. Description: Uses data from administrative records of government agencies. STUDY DESCRIPTION & STRENGTHS OF APPROACH Not always possible to discriminate between different causal effects WEAKNESSES/ LIMITATIONS OF APPROACH Uses descriptive research to identify associations between indicators related to social exclusion. Measures the direct association between an area characteristic and an outcome measure, controlling for individual characteristics that may also influence that outcome. HOW WEAKNESSES HAVE BEEN ADDRESSED Significant associations are found between neighbourhood characteristics and nonhousing outcomes. Area is an important influence, but there are equally and more important influences at the individual and household levels. Evidence that in deprived areas people’s expectations of starting a job and actually starting a job are lower than in non-deprived areas. The paper draws a linkage between core housing need and labour • • • • 63 The Jobs-Plus program increased the employment and earnings of residents in public housing developments relative to the comparison group. This impact was sustained over time. • PROVEN LINKAGES & CONCLUSIONS Erebus International. 2005. Review of the Recent Literature on Socio-economic Status and Learning Acceptable Housing: Canadians in Core Housing Need STUDY Literature review. METHODOLOGY Socio-economic status intersects with other factors including gender, culture and ethnicity. Some indications support the view that it is the concentration of disadvantage versus disadvantage itself that is significant in educational underperformance. Parent level of education and employment status are key determinants of advantage/disadvantage • • • Strengths: Examines the results of the Longitudinal Surveys of Australian youth (LSAY) as well as international studies. 64 Results indicate that there continues to be an inter-relationship between education and socio-economic status. Strengths: Draws on data derived from the 2001 Census to explore the nature of the difficulties faced by Canadians in core housing needs. • force ties. PROVEN LINKAGES & CONCLUSIONS Description: Literature review to determine if the relationship between socio-economic disadvantage and learning outcomes holds in 2005. HOW WEAKNESSES HAVE BEEN ADDRESSED Households with weak ties to the labour force have lower incomes, are more likely to rent and are much more likely to be in core housing need than other households. WEAKNESSES/ LIMITATIONS OF APPROACH • relationship to a number of social issues. STUDY DESCRIPTION & STRENGTHS OF APPROACH METHODOLOGY Literature review. Literature review. STUDY Jencks, C. & Mayer, S. E. (1990). The Social Consequences of Growing Up in a Poor Neighbourhood In . Inner City Poverty in the United States Edwards, B. (2005). Does it Take a Village? An Description: Studies the effects neighbourhoods have on young children. Strengths: Focus on quantitative studies that try to separate neighbourhood or school effects from family effects through statistical analysis of survey data. Description: Compares children from similar families who grow up in different kinds of neighbourhoods. Looked at educational attainment and economic success as well as other outcomes. STUDY DESCRIPTION & STRENGTHS OF APPROACH Selection bias based on socio-demographic factors such as family Sample bias, random sampling errors, measurement error, and specification error were evident in a number of studies making it difficult to draw general conclusions. WEAKNESSES/ LIMITATIONS OF APPROACH Uses child and sociodemographic data as controls to limit the HOW WEAKNESSES HAVE BEEN ADDRESSED Neighbourhoods that are predominantly black or that have a high degree of welfare dependency reduce young men’s chances of having a high-paying job in adulthood. The greater the control for family background, the smaller the impact of neighbourhood or school effects. Findings from the literature review suggest that neighbourhoods • • • 65 The authors conclude that studies of schools and studies of neighbourhoods yield contradictory conclusions about the determinants of educational achievement. • and are more reliable predictors of educational outcomes than family or household income. PROVEN LINKAGES & CONCLUSIONS Jacob, B. A. (2003). Public Housing, Housing Vouchers and Student Achievement: Evidence from Public Housing demolitions in Chicago investigation of Neighbourhood Effects on Australian Children’s Development STUDY Experimental study. METHODOLOGY Strengths: Students are matched to housing developments through home addresses in school records and building closure is determined from occupancy data. Students living in CHA units slated for closure are matched with peers living in units in the same project that were not closed. Description: Uses administrative data from the Chicago Housing Authority (CHA) and the Chicago Public Schools. Strengths: Describes the Longitudinal Study of Australian Children (LSAC) designed to examine neighbourhood effects on children. The study uses a wide range of variables. STUDY DESCRIPTION & STRENGTHS OF APPROACH HOW WEAKNESSES HAVE BEEN ADDRESSED likelihood that neighbourhood influences were the result of selection bias. Uses demolition as an instrument to estimate the causal impact of living in public housing on educational outcomes. WEAKNESSES/ LIMITATIONS OF APPROACH income that may be associated with parents’ decisions to live in a particular neighbourhood. Can’t directly replicate the “treatment-on-thetreated” effect. It was also found that boys are more affected by neighbourhood disadvantage than girls. Low-income families tend to relocate close to their original neighbourhood. Students who moved were still living in high poverty neighbourhoods and attending schools identical to the control group. There was little to no impact on the academic achievement of these children. Study concludes that providing households the option of using housing • • • 66 The LSAC study found that children living in the two most advantaged neighbourhoods had significantly higher Learning domain scores than the other three neighbourhoods. • matter to children’s development. PROVEN LINKAGES & CONCLUSIONS Multivariate Kohen, E. Description: Study uses the Strengths: Experimental study design includes a control group, a “Section 8” (traditional voucher) group and an experimental voucher group. Empirical approach was to separately compare each treatment group to the control group on a wide range of measures. Survey data were collected, a sample of youth and an adult from the youth’s household were interviewed Description: Compares groups of youth involved in the Moving to Opportunity program. STUDY DESCRIPTION & STRENGTHS OF APPROACH To overcome these problems, Bonferroni20 and Bonferroni-Holm21 adjustments are used There is the potential of a familywise error rate (FWER)19 where the significance of treatment effects is viewed as a member of a family of hypotheses and per-comparison errors (coefficients viewed in isolation). Used multivariate HOW WEAKNESSES HAVE BEEN ADDRESSED WEAKNESSES/ LIMITATIONS OF APPROACH Females in the experimental group experienced improvements in education. The mental health of females in both voucher groups improved. Males in both treatment groups were more likely than controls to engage in risky behaviours and to experience physical health problems. “Results indicate that • • • • vouchers to relocate to different neighbourhoods will not necessarily produce better (or worse) educational outcomes for poor children. PROVEN LINKAGES & CONCLUSIONS 67 The familywise error rate is “the chance of any false positives is the standard measure of Type I errors in multiple testing” (Nichols & Hayasaka, 2003) 20 “The Bonferroni correction is a mathematical correction originally utilized to reduce falsely significant results in statistical analyses.” (source: Royal College of Surgeons of Edinburgh , 2007) 21 Variation of the above correction. 19 Experimental study. METHODOLOGY Kling, J. R. & Liebman, J.B. (2004). Experimental Analysis of Neighborhood Effects on Youth STUDY Hertzman, C. & Brooks-Gunn, J. (1998). Neighbourhood Influences STUDY analyses of longitudinal survey data. METHODOLOGY Strengths: focuses on specific areas of competency associated with children's school readiness in two age groups of children National Longitudinal Survey of Children and Youth (NLSCY) STUDY DESCRIPTION & STRENGTHS OF APPROACH WEAKNESSES/ LIMITATIONS OF APPROACH HOW WEAKNESSES HAVE BEEN ADDRESSED analyses to determine the effects of all variables on the responses of interest. Neighbourhood characteristics have an impact on children's readiness to learn. As would be expected for both toddlers and preschoolers, family characteristics are also important; they mediate neighbourhood effects for the youngest children. However, these effects remain significant over and above family characteristics for the older group of children • 68 Family characteristics such as higher levels of household income and higher levels of maternal education are also associated with children's competencies. • neighbourhood characteristics, particularly neighbourhood affluence and cohesion are associated with competencies for children of both age groups. PROVEN LINKAGES & CONCLUSIONS METHODOLOGY Empirical study. Empirical study Longitudinal study of administrative data STUDY Mullins, P. & Western, J. (2001). An Examination of the Relationships Between Housing Systems and Nonhousing Outcomes Oreopoulos, Phillip (2003). The LongRun Consequences of Living in a Poor Neighbourhood Concerns for some, small, self-selection bias of families in the housing allocation system. Potential for missing information on some tenants in using the IID. Strengths: Uses public housing data, matched to postal addresses, and tax Non-longitudinal study, therefore outcomes before and after housing assistance was given can’t be measured. WEAKNESSES/ LIMITATIONS OF APPROACH Description: Longitudinal study of Toronto public housing tenants examining neighbourhood effects on labour market outcomes. Strengths: Uses a cross sectional (comparative) analysis of different tenure groups according to nonhousing outcomes. Description: Uses data collected in 1997 as part of the South East Queensland Quality of Life Survey. STUDY DESCRIPTION & STRENGTHS OF APPROACH IID data crossreferenced with data from the Longitudinal Administrative Database (LAD). Self-selection concerns controlled for using various resident groups. HOW WEAKNESSES HAVE BEEN ADDRESSED Private rental housing provides better outcomes than public housing; but in groups low-income groups, the receipt of government subsidy assistance or not, only marginally affected non-housing outcomes. Tenure type is linked to relative poverty and advantage or disadvantage. Little to no correlation between wide-ranging neighbourhood quality and young resident’s chances for long-run labour market success; Above findings despite substantial differences • • • • 69 Home ownership confers the most benefits in terms of educational achievement, employment and income. • (i.e., preschoolers).” PROVEN LINKAGES & CONCLUSIONS Decomposition analysis using longitudinal survey data. Analysis of survey results. Peck. B. (2001). The Poor Stay Poor and the Rich Stay METHODOLOGY Pebley, A. R., & Sastry, N. (2003). Concentrated Poverty vs. Concentrated Affluence: Effects on Neighborhood Social Environments and Children’s Outcomes STUDY Description: An Australian study that focuses on the associations between SES and Strengths: Sampled children were administered standardized tests and sampled children’s Primary Caregiver was interviewed. In addition, sampled children over 9 were interviewed as was one sibling Description: Uses data from the first wave of the longitudinal Los Angeles Family and Neighbourhood Survey (L.A.FANS). administrative data (Intergenerational Income Database- IID). Naturally occurring randomized selection situation, enabling examination of data over a period of about 30 years. STUDY DESCRIPTION & STRENGTHS OF APPROACH Variables of family earning, maternal education, neighbourhood income and maternal reading scores are interrelated and are also associated with other child, family, and neighbourhood characteristics. WEAKNESSES/ LIMITATIONS OF APPROACH Uses a decomposition analysis where socioeconomic inequalities in test scores are decomposed into two components related to test scores and socioeconomic variables. HOW WEAKNESSES HAVE BEEN ADDRESSED The study concluded that on average, compared with students • 70 Further research recommended on sibling differences. Neighbourhood level median income is an important predictor of children’s achievement, even when observable and unobservable family characteristics are held constant. • • Family differences matter a great deal, as measured using sibling outcome correlations. • in average household income, parental education attainment, family composition, parental welfare participation, and crime exposure. PROVEN LINKAGES & CONCLUSIONS Mixed methods study. Use of econometric Steele, Marion and Sarker, Rakhal METHODOLOGY Sanbonmatsu, L., Kling, J.R., Duncan, G.J., & BrooksGunn, J. (2004). Neighborhoods and academic achievement: Results from the moving to opportunity experiment Rich STUDY Description: Use of Statistics Canada microdata sets, Strengths: Looked at children ages 6 – 20. One adult and up to 2 children from each family were selected for the data collection. Description: Uses the randomized housing mobility experiment, Moving to Opportunity to estimate the causal effects on children’s educational outcomes of moving out of high-poverty neighbourhoods. Strengths: Data were collected through survey results. a range of educational indicators. STUDY DESCRIPTION & STRENGTHS OF APPROACH Presence of estimation problems. Since only Possibility of interviewer effects on the administration and scoring of test items. WEAKNESSES/ LIMITATIONS OF APPROACH Interviewer effects controlled for by computing adjusted test scores. Authors estimated the interviewer coefficient conditional on census tracts, fixed effects and individual characteristics, and calculated the estimated interviewer effect as the deviation of the interviewer from the site mean. Other investigators using econometric HOW WEAKNESSES HAVE BEEN ADDRESSED The experimental treatment led to a substantial shift in the poverty rank of the neighbourhoods children lived in, but a smaller change in the distribution of school ranks. Personal and family wellbeing, including • • 71 Achievement-related benefits for school-aged children from moving to improved neighbourhood environments are small. • from high SES families, those from low SES families leave school earlier, have lower aspirations, tend to have different patterns of subject selection in postcompulsory schooling, achieve at a lower level at school, obtain lower Tertiary Entrance Scores, are less likely to go to university, and are more likely to enter occupations associated with low SES PROVEN LINKAGES & CONCLUSIONS METHODOLOGY techniques using Statistics Canada microdata sets. Mixed methods study. STUDY (2005). Estimates of the Private and Societal Costs and Benefits of Homeownership in Selected Canadian Cities U.S. Department of Housing and Urban Development (2003). Moving to opportunity, interim impact evaluation. Strengths: Study uses both quantitative and qualitative data to assess MTO’s effects in areas including child educational achievement and adult and youth employment. MTO randomly assigned subjects to a variety of housing options. Description: Study designed to examine the impact of MTO halfway through the 10-year research period. including those from the General Social Surveys. STUDY DESCRIPTION & STRENGTHS OF APPROACH Control group has not been static. Many control group families moved out of public housing. In addition, the welfare system changes had effects on low-income families. In addition, there were increased opportunities for all MTO families because of a booming economy. This would have an impact on the results. WEAKNESSES/ LIMITATIONS OF APPROACH income and usually one other variable are controlled, it may be that what is identified as the ownership impact arises because of other variables which are not controlled for. Three problems are identified: co-linearity; self-selection; and endogeneity HOW WEAKNESSES HAVE BEEN ADDRESSED techniques to overcome these problems have found that estimation without using these techniques doesn’t change results much. Use pattern of results across different agegroups to gauge whether the problems are likely to matter. Use findings from international literature to corroborate findings. MTO had small but significant effects on the characteristics of the schools sample children attended. Therefore nearly 3/4s of the experimental children still attended schools in the same school district they were in at baseline. Results show that the demonstration had virtually no significant effects on any of the measures of educational performance analyzed. • • 72 Owner children do better than renter children on standardized tests and behaviour outcomes. • improved child outcomes is a benefit of homeownership. PROVEN LINKAGES & CONCLUSIONS List of Acronyms CHPC: Citizens Housing and Planning Council CMHC: Canada Mortgage and Housing Corporation HUD: U.S. Department of Housing and Urban Development LAD: Longitudinal Administrative Database LSIC: Longitudinal Survey of Immigrants to Canada MTO: Moving to Opportunity NLSCY: National Longitudinal Survey of Children and Youth PUMF: Public use microdata files RDC: Research Data Centre (Statistics Canada) SCPI: Supporting Community Partnerships Initiative SES: Socio-economic status SLID: Survey of Labour Income and Dynamics SUTVA: Stable unit treatment value assumption 73 Bibliography Allard, S. W, & Danziger, S. (2001). Proximity and opportunity: How residence and Race Affect the Employment of Welfare Recipients. Almey, M. (1995). Housing and Household Facilities. Women in Canada: A Statistical Report. 3rd Edition, Statistics Canada. Ottawa. Armstrong, A., Francis, R., Bourne, M., and I. Dussuyer (2002). Difficulties of Developing and Using Social Indicators to Evaluate Government Programs: A Critical Review. Paper presented at 2002 Australasian Evaluation Society International Conference. Baker, Michael (ed.) (2002). The Westburn Dictionary of Marketing. Westburn Publishers. Beauvais, Caroline and Jenson, Jane (2003). The Well-being of Children: Are There “Neighborhood Effects”? Discussion Paper. Begin, Patricia, Lyne Casavant, Nancy Miller Chenier, and Jean Dupuis, 1999. Homelessness. National Library of Parliament, Ottawa. Bloom, H.S., Riccio, J.A., Verma, N. (2005). Promoting work in public housing: The effectiveness of Jobs-Plus. MDRC. Blunden, H. with Johnston, C. (2005). Public Housing and Nonhousing Outcomes – A Background Paper. Shelter New South Wales, Australia. Bradford, Neil (2005). Place-based Public Policy: Towards a New Urban and Community Agenda for Canada. Research Report F51. Canadian Policy Research Networks. Braconi, F. (2001). Housing and Schooling. Citizens Housing and Planning Council: The URBAN PROSPECT. Housing Planning and Economic Development, Volume 7, Number 2, . Buck, Nick (2001). Identifying Neighborhood Effects on Social Exclusion. Journal of Urban Studies, Volume 38, Number 12, pp.2251-2275. Burke, T., and K. Hulse (2002). Sole Parents, Social Wellbeing and Housing Assistance. AHURI, Australia. Canadian Council on Social Development (2002). The Progress of Canada’s Children 2002. CCSD, Canada. 74 Canada Mortgage and Housing Corporation (2004). Transitional Housing: Objectives, Indicators of Success, and Outcomes. CMHC Research Highlight, Socio-economic Series, Number 04-017, February 2004. Canada Mortgage and Housing Corporation (2000). Women on the Rough Edge: A Decade of Change for Long-Term Homeless Women. CMHC Research Highlight, Socio-economic Series, Number 54, June 2000. Carter, Tom and Chesya Polevychok (2004). Housing is Good Social Policy. Canadian Policy Research Networks. Chiodo, D., Leschied, A.W., Whitehead, P.C. and D. Hurley (2003). The Characteristics of Abused Women in the Caseload of a Child Protection Service. The University of Western Ontario. Chisholm, S. (2002). Adjusting Housing Practice Towards a Social Inclusion Framework. Presentation to the CCSD Conference, March 2002. Chisholm, S. (2001). Housing and Social Inclusion: Asking the Right Questions. Laidlaw Conference Presentation, CHRA. Cooper, Merrill (2001). Housing Affordability: A Children’s Issue. Canadian Policy Research Networks Inc. Currie, J. & Yelowitz, A. (1999). Are Public Housing Projects Good for Kids? National Bureau of Economic Research. CurtisL. M. Dooley, and S. Phipps (2004). Child Well-Being and Neighbourhood Quality: Evidence From the Canadian National Longitudinal Survey of Children and Youth. Social Science & Medicine, Volume 58, Issue 10, pp.1917-1927. Curtis, L. and S. Phipps (2000). Economic Resources and Children’s Health and Success at School: An Analysis Using the NLSCY. Human Resources Development Canada. Dodson, J. (2005). Is There a Spatial Mismatch Between Housing Affordability and Employment Opportunity in Melbourne? Australian Housing and Urban Research Institute. Urban Policy Program, Griffith University. Douglas, D. (2005). Immigrant and Refugee Families: Challenges of Settlement. Better Together: Housing and Families Conference. Dunn, J.R. (2002). Housing and Inequalities in Health: A Study of Socioeconomic Dimensions of Housing and Self Reported Health from a 75 Survey of Vancouver Residents. Journal of Epidemiology and Community Health, Volume 56, Number 9, pp.671-681. Dunn, J.R. (2000). Housing and Health Inequalities: Review and Prospects for Research. Housing Studies, Volume 15, Number 3, pp.341-366. DTZ Consulting and Research for Department of Trade and Industry (2006). Housing, Economic Development and Productivity: Literature Review. Edwards, B. (2005). Does it Take a Village? An investigation of Neighbourhood Effects on Australian Children’s Development. Australian Institute of Family Studies. Family Matters. Number 72, Summer 2005, pp. 36-43. Ekos Research Associates Inc. (1998). The Use of Social Indicators as Evaluation Instruments. Human Resources Development Canada. Ellen, Ingrid and Turner, Margery (1997). Does Neighbourhood Matter? Assessing the Recent Evidence. Housing Policy Debate 8.4. Engeland, John, and Roger Lewis (2004). Exclusion from Acceptable Housing: Canadians in Core Housing Need. Horizons, 7.2. Policy Research Initiative, pp.27-33. Erebus International. (2005). Review of the Recent Literature on Socioeconomic Status and Learning. Report to the NSW Department of Education and Training. Australia. Evans, G.W., Wells, N.M., A. Moch (2003). Housing and Mental Health: A Review of the Evidence and a Methodological and Conceptual Critique. Journal of Social Issues, Volume 59, Number 3. Blackwell Publishing, pp.475500. Fallis, George (1994). “The Social Policy Challenge and Social Housing” in Home Remedies: Rethinking Canada’s Housing Policies (George Fallis et al. Eds). CD Howe Institute. Galster G., and S. Killen (1995). The Geography of Metropolitan Opportunity: A Reconnaissance and Conceptual Framework. Housing Policy Debate, Volume 6, issue 1. Fannie Mae, pp. 7-43. Garson, D.G. (2006). Ethnography, in Statnotes: Topics in Multivariate Analysis. North Carolina State University, retrieved 06-25-2008, http://www2.chass.ncsu.edu/garson/PA765/statnote.htm 76 Godwin, Arku, 2005. The Evolution of Ideas about the Relationship between Housing and Economic Development: Ghanaian Policy in an International Context. McMaster University, Hamilton, Paper AAINR04214. Government of Canada (2004). Creating a National Seniors Agenda: Report of the Prime Minister’s Task Force on Active Living and Dignity for Seniors. Graham, K. and E. Peters (2002). Aboriginal Communities and Urban Sustainability. Canadian Policy Research Networks. Grant, G. F. (2001). Housing Mobility and Mixed-Income Housing: An Assessment of Current Urban Poverty Deconcentration Strategies. Kennedy School Review. Boston, MA. Hay, David (2005). Housing, Horizontality and Social Policy. Canadian Policy Research Networks. Canada. Health Canada (2007). People, Place and Health. Health Policy Research Bulletin, Issue 14, November 2007. Hertzman, C. and C. Power (2006). A Life Course Approach to Health and Human Development in Jody Heyman, et al. (ed.), Healthier Societies: From Analysis to Action. Oxford University Press, pp. 83-106. Heckman, J. (1979). Sample Selection Bias as a Specification Error. Econometrica. Volume 47, Number 1, pp.153-161. Hulchanski, D. (2005) Rethinking Canada’s Housing Affordability Challenge. University of Toronto - Centre for Urban and Community Studies, Canada. Hulchanski, D. (2005b) No Homeland for the Poor: Houselessness and Canada’s Unhoused Population. Canadian Conference on Homelessness, Toronto, May 2005. Plenary Session Presentation. Hulchanski, D. (1997). Immigrants and Access to Housing: How Welcome are Newcomers to Canada? Presentation to Metropolis Year II Conference. Hulse, K. & Randolph, B. (2004). Getting a Job: The Role of Rental Housing Assistance. European Network for Housing Research, Housing: Growth and Regeneration, Cambridge (UK). Hwang, S. (2001). Homelessness and Health. Canadian Medical Association Journal; 164 (2), pp. 229-233 77 Jackson, A. (2004). Home Truths: Why the Housing System Matters to All Canadians. Ottawa: Canadian Centre for Policy Alternatives and the Canadian Housing Renewal Association. Jencks, C. & Mayer, S. E. (1990). The Social Consequences of Growing Up in a Poor Neighbourhood. Inner City Poverty in the United States. L. Lynn and M.G.H. McGeary (Eds), p. 111-186. Washington, DC: National Academy Press. Johnson, L.C. & Ruddock, A. (2000). Building Capacity: Enhancing Women’s Economic Participation Through Housing. Canadian Housing and Renewal Association. Canada. Josephson, Gordon (2004). Relating Homelessness to Education, Employment and Income Support: A Review of Canadian Literature. Horizons, 7(2). Policy Research Initiative, pp.66-71. Kling, J. R. & Liebman, J.B. (2004). Experimental Analysis of Neighborhood Effects on Youth. National Bureau of Economic Research. Kramer, F. D. (2000). Designing a Family-Centered, Housing-Based Employment Program. Welfare Information Network, Volume , Issue 1, April. Kohen, E. Hertzman, C. & Brooks-Gunn, J. (1998). Neighbourhood Influences. Human Resources and Social Development Canada. Kohen, D. E., Hertzman, C. (1999). Neighbourhood affluence and School Readiness. Education Quarterly Review. Volume 6, Number 1, pp. 44-51. Labonte, R. and R. Huckfeldt (1999). Social Capital and Community Development. Australian and New Zealand Journal of Public Health, Volume 23, pp. 430-433. Levy, D.K., & Woolley, M. (2007). Relocation is Not Enough. Employment Barriers Among HOPE VI Families. Urban Institute. Washington DC. Maclennan, Duncan (2001). Putting Housing in its Place. Presentation to Australian National Housing Conference. Manski, C. (1993). The Identification of Endogenous Social Effects: The Reflection Problem. Review of Economic Studies, 60 (3), pp. 531-542. May, Jeff (2007). Social Lives in Social Housing: Resident Connections to Social Services. Canadian Policy Research Networks, Ottawa. Mendelson, Michael (2006). Building Assets Through Housing. Caledon Institute of Social Policy, Ottawa. 78 Mendez, P., Hiebert, D., 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, Volume 15, Issue 2, supplement pp. 82-104. Ministry of Social Development (2004). The Social Report. Ministry of Social Development of New Zealand. Mullins, P., Western, J., and Broadbent, B. (2001). The Links Between Housing and Nine Key Socio Cultural Factors: A Review of the Evidence. AHURI Positioning Paper, Melbourne. Mullins, P. & Western, J. (2001a). An Examination of the Relationships Between Housing Systems and Non-housing Outcomes. National Housing Conference, Brisbane, Australia. Mullins, P., and Western, J. (2001b). The Links Between Housing and Nine Key Socio Cultural Factors. Australian Housing and Urban Research Institute Final Report 5. National Association of Friendship Centres (NAFC) (2004). Towards an Affordable Housing Strategy for Urban Aboriginal Peoples. NAFC, Canada. Neill, James (2006). Quantitative Research Methods: Meta-Analysis Research Methodology. http://wilderdom.com/research/meta-analysis.html. Nichols, T. and S. Hayasaka (2003). Controlling the Familywise Error Rate in Functional Neuroimaging: A Comparative Review in Stat Methods Med Res, Volume 12, Number 5 (October 2003), pp. 419-446. Oreopoulos, Phillip (2003). The Long-Run Consequences of Living in a Poor Neighborhood, Quarterly Journal of Economics. Pebley, A. R., & Sastry, N. (2003). Concentrated Poverty vs. Concentrated Affluence: Effects on Neighborhood Social Environments and Children’s Outcomes. RAND Labor and Population, Working Paper Series 03-24. California. Peck. B. (2001). The Poor Stay Poor and the Rich Stay Rich. Issues in Educational Research, vol. 11, http://education.curtin.edu.au/iier/iier11/peck.html Phibbs, Peter and Peter Young (2005). Housing Assistance and Non-Shelter Outcomes, Final Report. Australian Housing and Urban Research Institute. 79 Phibbs, Peter with Peter Young (2002). Housing Assistance and Non-Shelter Outcomes. Australian Housing and Urban Research Institute, Sydney Research Centre. Policy Research Initiative (PRI) (2005). Housing Policy and Practice in the Context of Poverty and Exclusion: Synthesis Report. PRI Project, Canada. Policy Research Initiative (PRI) (2004). A Life-Course Approach to Social Policy Analysis: A Proposed Framework. PRI Project, Canada. Pomeroy, S. (2004). Leaks in the Roof, Cracks in the Floor: Identifying Gaps in Canada’s Housing System. Canadian Housing and Renewal Association (CHRA), Canada. Pope, J. (2003). Social Capital and Social Capital Indicators: A Reading List. Public Health Information Development Unit, Australia. Popkin, S.J., Buron, L.F., Levy, D.K., and M.K. Cunningham (2000). The Gautreaux Legacy: What Might Mixed-Income and Dispersal Strategies Mean for the Poorest Public Housing Tenants? Housing Policy Debate. Volume 11, Number 4, pp.911-942. Popkin, S.J., Leventhal, T., & Weismann, G. (2006). Girls in the ‘hood: Evidence on the impact of safety. Poverty 7 Race Research Action Council. Sept/Oct. 2006 Newsletter. Renaud, J., and K. Bégin (2006). The Residential Mobility of Newcomers to Canada: The First Months. Canadian Journal of Urban Research, Volume 15, Issue 2, supplement pp. 67-81. Riccio, J.A. (2007). Subsidized housing and employment: Building evidence about what works to improve self-sufficiency. Joint Center for Housing Studies, Harvard University. Rohe, William, Shannon Van Zant, and George McCarthy (2001). The Social Benefits and Costs of Homeownership: A Critical Assessment of the Research. Joint Center for Housing Studies, Harvard University. Rosenbaum, J.E., Reynolds, L., & Deluca, S. (2002). How do places matter? The geography of opportunity, self-efficacy and a look inside the black box of residential mobility. Housing Studies 17.1. Ross, N., Wolfson, M., Dunn, J., Berthelot, J-M., Kaplan, G, and J. Lynch (2006). Relation Between Income Inequality and Mortality in Canada and in the United-States: Cross Sectional Assessment Using Census Data and Vital Statistics. British Medical Journal, Volume 320, pp.898-902. 80 Royal College of Surgeons of Edinburgh (2008). Statistics Definitions accessed at http://www.edu.rcsed.ac.uk/statistics/the%20bonferroni%20correction.htm. Sampson, R., Morenhoff, J.D., and T. Gannon-Rowley (2002). Assessing “Neighbourhood Effects”: Social Processes and New Directions in Research. Annual Review of Sociology, 28, pp. 443-478. Sanbonmatsu, L., Kling, J.R., Duncan, G.J., & Brooks-Gunn, J. (2004). Neighborhoods and academic achievement: Results from the moving to opportunity experiment. National Bureau of Economic Research (NBER) Princeton University. Shlay, A. (1995). Housing in the broader context in the United States. Housing Policy Debate. Volume 6, Issue 3, Fannie Mae, pp. 695-720. SHS Consulting (2007). Literature Review and Assessment: Linkages Between Housing and Non-Housing Outcomes Related to Education, Skills Development and Employment. Phase One: Overview of the Current State of Knowledge on Societal Outcomes of Housing, CMHC. Canada Smith, R., and S. Torjman (2004). Policy Development and Implementation in Complex Files – Lessons from the National Homelessness Initiative – Lessons from “Vibrant Communities”. Canada School of Public Service, Canada. Sobel, M. (2006). What Do Randomized Studies of Housing Mobility Demonstrate?: Causal Inference in the Face of Interference. Journal of the American Statistical Association, Volume 101, Number 476, pp. 1398-1407. Social Housing Services Corporation (2007). Snakes and Ladders: Ending Poverty Traps by Rebuilding Livelihoods in Social Housing. Toronto. Statistics Canada (2006). Aboriginal Peoples in Canada in 2006: Inuit, Métis, and First Nations, 2006 Census. Catalogue number 97-558-XIE. Ottawa, Statistics, Canada. Statistics Canada (2004). Neighbourhood Inequality, Relative Depravation and Self-perceived Health Status. Catalogue number 11F0019MIE No.228. Ottawa, Statistics Canada. Statistics Canada (2003). Neighbourhood Low Income, Income Inequality and Health in Toronto. Health Reports, Statistics Canada, Catalogue number 82-003. Volume 14, Number 2, pp.21-34. 81 Statistics Canada (2007). 2006 Census: Community Profiles – Canada (table). Ottawa , Statistics Canada, Catalogue no. 92-591-XWE. Statistics Canada (2002). 2001 Census: Community Profiles – Canada. Ottawa, Statistics Canada, Catalogue no. 93F0053XIE. Statistics Canada (1996). 1996 Census: Private Households by Household Type, Showing Number of Household Maintainers – Canada. Ottawa, Statistics Canada, Nation Series, Catalogue no. 93F0030XDB96000. Steele, Marion (2008). Memorandum on Housing Research. Sent to SHS Consulting. Steele, Marion and Sarker, Rakhal (2005). Estimates of the Private and Societal Costs and Benefits of Homeownership in Selected Canadian Cities. CMHC. TD Economics Special Report (2003). Affordable Housing in Canada: In Search of a New Paradigm. Tremblay, S., Ross, N. and J-M Berthelot (2001). Factors Affecting Grade 3 Student Performance in Ontario: A Multilevel Analysis. Statistics Canada Education Quarterly Review. Volume 7, Number 4, pp. 25-41. U.S. Department of Housing and Urban Development (2003). Moving to opportunity, interim impact evaluation. National Bureau of Economic Research. Varady, D., and C. Walker (2003). Housing Vouchers and Residential Mobility. Journal of Planning Literature, Volume 18, Number 1, pp. 17-30. Von Hoffman, A., Belsky, E., and K. Lee (2006). The Impact of Housing on Community: A Review of Scholarly Theories and Empirical Research. Joint Center for Housing Studies, Harvard University. Wilkinson, R. and M. Marmont (Ed.) (2003). Social Determinants of Health Second Edition: The Solid Facts. Denmark: World Health Organization. Williams, S. (2003). From HOPE VI to HOPE Sick? Boston: Dollars and Sense, July 2003. Wilson, W.J. (1987). The Truly disadvantaged. Chicago: University of Chicago Press. 82
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
advertisement
© 2021