An approach to modeling and forecasting real estate residential property market

Abstract

This thesis aims to provide an approach to real estate residential modeling and forecasting covering property types’ correlation, time series attributes within a region or a city, and socio-economic attributes of preferred real estate locations. The thesis covers residential estate markets and concentrates on property types, while previous studies that have considered country wide house price indices. There is a gap identified in the literature in the need to study correlations between property types within a region or a city and whether they will provide diversification benefits for real estate investors such as risk reduction per unit of returns. The thesis concentrates on property type seasonality in addition to modeling time series attributes within a region or city instead of real estate index seasonality. This thesis the first to combine modern information systems techniques such as geographic information systems (GIS) with socio-economic factors to help understanding causal relationships that can be used to forecast real estate prices. The results show that it is more achievable to forecast real estate prices within a city than for the real estate market of the entire country. The GIS and socio-economic modeling results show that higher property prices are awarded to real estate with more green spaces, residents with higher disposable incomes, lower council tax bands, fewer tax benefits claimants, and better health services. Previous studies have examined real estate price indices at the macro level (the general, all real estate house price indices). There has not been a study that examines real estate price forecasts by property types within a city. The contribution of this thesis is its focus on time series analysis as well as causal modeling within a city with the objective of providing a better understanding of the dynamics of real estate price changes.