Ripple effects in the housing market in the Greater Bay Area, China: A spatial - temporal analysis using different spatial weights matrices

Regional integration and urbanization are the two intertwining trends that are shaping the economic geography landscape of cities. During this process, the dynamic intercity relationships within urban clusters has important impacts on regional housing markets. In the case of GBA, various types of spatial interaction have been forming a network structure, which provided possibilities for close linkages across housing markets. Therefore, the spatial-temporal patterns of ripple effects are potentially one of the most significant characteristics in understanding the regional housing markets and urban agglomeration as a whole. The thesis adopts the GBA as a case to examine the ripple effects from the different channels of regional integration, and tests how the interface of intra-city socio-economic fundamentals and inter-city relationships function in ripple effects, that is influencing the housing price of cities itself as well as the spill-over effect. The analysis of the thesis was developed in three aspects. Firstly, it evaluated the characteristics of ripple effect in terms of short-term volatility, spatial aggregation and diffusion, and spatial lead-lag relationships. This is the fundamental validation of ripple effects as spatial-temporal patterns of housing prices across markets. Secondly, it described the patterns and compared relationships of multi-dimensional regional integration as the potential transmission channels of ripple effect. The five facets of regional integration are geographical contiguity, transportation accessibility, economic gravity, institutional proximity and information transmission. The last part explores the mechanisms of how multi-dimensional integration functions as the transmission channels of the ripple effect by estimation in dynamic spatial models. In this context, geographical adjacency is not necessarily associated with spatial diffusion of housing price in this case. Instead, spatial linkages other than physical distance are considered as structuring elements in the spatial framework for ripple effects. However, conclusion verified that geographical distance is still the physical condition that influences the formation of other spatial linkages. But its role has becoming more and more indirect and subtle, as it is not the single determination for spatial leading-lag relationships. As for the impacts of regional integration on ripple effect. The results of dynamic Spatial Durbin Models indicated that the general direction of ripple effects is threefold: positive spatial diffusion effects, positive time lag effects and negative spatial-temporal diffusion effects. However, the magnitudes are different across matrices. Information transmission is the most prominent channel for spatial diffusion and time lag effects, while transportation accessibility is more significant for the negative spatial-temporal diffusion effects. Also, common factors are considered to explain the role of general markets situations, while marginal effects are further analysed to test the the direct and indirect effect of socio-economic fundamentals in both the short- and long- run. In terms of policy recommendation, it is to design a regional management mechanism that the multi-dimensions relationships in a comprehensive and dynamic way, instead of limited on the static geographical profiles. Also, it is crucial for each city to reconsider and establish its general and sector strategy under the framework of regional development. Keywords: Ripple effect, regional integration, housing price, dynamic Spatial Durbin Model