Use this URL to cite or link to this record in EThOS:
Title: Methodological and empirical challenges in modelling residential location choices
Author: Zolfaghari, Alireza
ISNI:       0000 0004 2752 7187
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Access from Institution:
The modelling of residential locations is a key element in land use and transport planning. There are significant empirical and methodological challenges inherent in such modelling, however, despite recent advances both in the availability of spatial datasets and in computational and choice modelling techniques. One of the most important of these challenges concerns spatial aggregation. The housing market is characterised by the fact that it offers spatially and functionally heterogeneous products; as a result, if residential alternatives are represented as aggregated spatial units (as in conventional residential location models), the variability of dwelling attributes is lost, which may limit the predictive ability and policy sensitivity of the model. This thesis presents a modelling framework for residential location choice that addresses three key challenges: (i) the development of models at the dwelling-unit level, (ii) the treatment of spatial structure effects in such dwelling-unit level models, and (iii) problems associated with estimation in such modelling frameworks in the absence of disaggregated dwelling unit supply data. The proposed framework is applied to the residential location choice context in London. Another important challenge in the modelling of residential locations is the choice set formation problem. Most models of residential location choices have been developed based on the assumption that households consider all available alternatives when they are making location choices. Due the high search costs associated with the housing market, however, and the limited capacity of households to process information, the validity of this assumption has been an on-going debate among researchers. There have been some attempts in the literature to incorporate the cognitive capacities of households within discrete choice models of residential location: for instance, by modelling households’ choice sets exogenously based on simplifying assumptions regarding their spatial search behaviour (e.g., an anchor-based search strategy) and their characteristics. By undertaking an empirical comparison of alternative models within the context of residential location choice in the Greater London area this thesis investigates the feasibility and practicality of applying deterministic choice set formation approaches to capture the underlying search process of households. The thesis also investigates the uncertainty of choice sets in residential location choice modelling and proposes a simplified probabilistic choice set formation approach to model choice sets and choices simultaneously. The dwelling-level modelling framework proposed in this research is practice-ready and can be used to estimate residential location choice models at the level of dwelling units without requiring independent and disaggregated dwelling supply data. The empirical comparison of alternative exogenous choice set formation approaches provides a guideline for modellers and land use planners to avoid inappropriate choice set formation approaches in practice. Finally, the proposed simplified choice set formation model can be applied to model the behaviour of households in online real estate environments.
Supervisor: Sivakumar, Aruna ; Polak, John Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral