Fringe area growth in Metropolitan Portland: an analysis of space-time variations in residential housing and land conversion, 1970-1980

摘要

Urban ecological problems have hitherto been addressed using one of two major approaches. The first has a social impetus directed at ethnic, economic and family characteristics and their relationships with the spatial distribution of urban residential housing. The second approach emphasizes the influence of the physical environment and the services available to subareas. The sociological method has had much more attention in modeling applications than the physical analytic technique. This study adopts the physical approach with a focus that is emphatic on infrastructural factors and land attributes, and their influence on the differential rates of fringe area residential growth in the Portland metropolitan region. Data were acquired through direct research supplemented by building permit records, jurisdictional estimates, and information from the 1970 and 1980 u.s. Censuses. Growth functional relationships were operationalized using housing starts and residential land conversion as two dependent variables against which the explanatory factors of infrastructure (water and sewer), land characteristics, road network density, accessibility, and social factors were regressed in recursive models over three subperiods in the decade 1970-1980. Five models were derived for the SMSA and the four counties for the decade, and three more subperiodic models for each area, for the categories of housing starts and land conversion. The derived models were tested against a standard econometric technique (Chow test) to verify the consistency of the coefficients (elasticities) over the different subareas in the four time periods. The results showed extremely high levels of significance of the Chow tests, deeming it necessary to examine the behavior of the elasticities in more detail over space and time. The results of the examination verified that the performance of infrastructure variables were highest in Washington County, while accessibility and road network density showed very high performances in Multnomah County. Land attributes were most notable in Clark County, while income elasticities were equally high in Multnomah, Washington, and Clark Counties. The lag effects of residential development in the immediate anteceding period were more important in Multnomah and Washington than in other counties. In Clark County, residential development in the early part of the decade was the only significant lag variable in models of the latter part of the decade. The conducted tests lend adequate support to the postulated hypotheses. In general, there was differential response to the selected attributes in the subareal models. Also, the results and tests confirmed that parameter estimates of attributes varied in different governmental jurisdictions. This implies that the counties placed different emphasis on the tested variables. Where the favorable set of variables was emphasized with one major sewer service district (Washington County), fringe area growth was enhanced. The emphasis of congestion-related variables (Multnomah County) without the desired infrastructure resulted in a relatively reasonable decline in fringe area residential housing.