Development of an operational system for monitoring the changes in urban subarea residential housing status: a spatial analytic application of the formulations of neighborhood filtering and neighborhood dynamics

Abstract

This dissertation focuses on the precise and meaningful measurement of change as it pertains to the urban subarea residential housing status. The word "measurement" is qualified as meaningful in the sense that the approach adopted is of relevance to public policy. Specifically, the dissertation is aimed at providing answers to the following research questions: First, how can the changes in residential housing status in the different parts of an urban area be precisely and meaningfully measured? Second, what variables are most appropriate for the measurement? Third, can these variables be useful for differentiating between various parts of the urban area? Fourth, do the resul ts of an urban subarea housing classification system depend on the specific variables used in the classification? Using data drawn mainly from the 1960 and 1970 censuses of housing for Portland, Oregon SMSA, a simple but robust methodology is developed for indexing and monitoring changes in the urban subarea residential housing status. The research borrows appreciably from Fisher and Winnick's, and Toulan' s formulations of the filtering process in the urban housing market. The variables used in the measurement and classification analyses include the changes in the following variables: median home value or contract rent, median household income relative to the average household size, housing quality, percentage of all occupied housing units, and percentage of owner occupied housing units. Principal component analysis is used for construction of composite index of change in urban subarea residential housing status. Furthermore, this composite index is used in a multivariate linear discriminant analysis for the classification of the various subareas (census tracts) in Portland, Oregon SMSA. The findings validate the variables employed in the analyses, and support the hypothesis that the results of an urban subarea classification system depend, to some extent, on the housing market variables used in the classification. The findings from the study show that operationally simple but robust systems can be developed for monitoring the changes in residential housing status in urban neighborhoods, in relation to the general urban area.