An exploration of the efficacy of Muth’s Access Space Trade-Off Model using Geographically Weighted Regression to model the Glasgow city region housing market
Dissertation Research Paper Technical Report and Supporting Document.docx (8.275Mb)
MetadataShow full item record
The Access Space Trade-Off Model (ASTOM) developed by Richard Muth (1969) has remained since its publication, the dominant and most researched theory of residential housing market formation. This paper used a Hedonic Price Model, implemented using Geographically Weighted Regression (GWR) to test the efficacy of Muth’s Model within the context of the Glasgow city region. The results of the Hedonic Price Model implemented via GWR show it to be an effective method for isolating the value different attributes of housing with R2 values showing model fit to vary between 0.73 and 0.83. Coefficients values were extracted from the GWR results to isolate the impact that location (distance from Glasgow city centre) has on house price, to understand whether this conforms to the expectations of the ASTOM. The ASTOM predicts a negative exponential gradient in house price per square meter measuring from the city centre. Location coefficients, were used as a proxy for house price per square metre, testing their impact on price in isolation from the other housing attributes. The results of this analysis show that when other attributes were controlled for, distance from the city centre generally has a positive impact on price. Neighbourhood characteristics play a more significant role in price, as do internal characteristics of the properties. There was no consistent negative price gradient when location coefficients were plotted against distance from the city centre. Improving model fit by adding richer data reduced fit of the results to the ASTOM.