The aim of thesis is to analyse the economic model of urban residential location through the application of duality methods. Whilst some dual methods have been used in urban economic modelling in the past this paper proposes alternative dual approaches which appear to be novel, but are complimentary to existing approaches to the urban model. As part of the application of dual techniques the paper proposes a method of application which is general enough to be applied to all Von Thunen type models and tests this proposition on the fundamental agrarian model of Von Thunen. As part of the dual analysis of the urban residential model the conditions for the traditional lot size hypothesis are examined in the light of conditional demand functions stemming from the dual analysis. The work also empirically tests the traditional residential lot size hypothesis for various Australian cities. The empirical method adopted involves estimation of density gradients utilising competing non-nested flexible form models and discrimination between these alternative models utilising semi-parametric non-nested tests based on an artificial regression model. Two of the three competing models have not been used in this context before, one of them being completely novel. Moreover, the artificial regression model has not been previously used in this context, requiring some modification to deal with the problem of competing models with dependent variable transformation. / Doctor of Philosophy (PhD)

Thesis (Ph.D.) - University of Queensland, 2005. / Includes bibliography.