Dynamic modeling of regional house price diffusion in Taiwan

This paper examines the lead-lag relationships and the dynamic linkages among four regional house price indices in Taiwan. We employ the Johansen cointegration technique, Toda and Yamamoto’s Granger causality test, the generalized impulse response approach, and variance decomposition analysis to find out the extent and the magnitude of their relationships. The estimated long-run relationship between regional house prices appears to have remained stable throughout the sample period. Our empirical results show a bidirectional relationship between house prices in the most important economic center, Taipei City, and its suburban area, Taipei County. However, there are no causalities of house prices between Taipei City and other megacities in Taiwan. The mutual impacts of the shocks between house prices in Taipei City and Taipei County are significantly positive, while these impacts on Kaohsiung City, far from Taipei City, are insignificant. Finally, the results of the generalized impulse response approach indicate that the house prices indices of Taipei City are the most exogenous while those for Taipei County are the most endogenous.