区域住房市场价格相关性研究——京津冀地区的实证分析-research on the correlation between regional housing market prices - an empirical analysis of beijing - tianjin - hebei region.docxVIP

PAGE PAGE VI Abstract Abstract Nowadays the whole society has paid more attention to the housing price volatility. Institutionally, some cities, especially in some areas of central city, would play a role as leader, which is so-called ＂ripple effect‖ in the academic. When modeling the housing price, we should not ignore the effect if some region exist ripple effect. This paper begin with the housing price model in order to introduce the ripple effect, and then present the probable cause of the ripple effect with China’s national conditions. Beijing Ttianjian and Hebei Province plays a key role at the economic circle around Bohai Sea. The region is closely related with politics, economy and culture, and its real market has been attention by people. However, there is little literature refer to real estate market of Beijing-Tinjin-Hebei. In this case, the paper discusses three hypotheses as follow: Hypothesis 1: the region could be regarded as a whole market, housing price converge in the long run. Hypothesis 2: as the largest economic entity, Beijing plays a central role in this region, which has influence over Tianjin and Hebei Province. Hypothesis 3: in the short turn, the housing price of the three markets whether has relative relation or independent of each other. In order to verify the first and the second hypotheses, we employed Granger Cointegration Test and Granger Causality Test. The result of Conintegration Tests manifested a fact that the housing prices in Beijing, Tianjin and Hebei has long-term convergence, which is, the price of those three places have the same trend of rise and fall. From the Granger Causality Test, the trend of housing prices fluctuation in Beijing is ahead of Tianjin and Hebei, so that we can believe that Beijing is in the center of the region, however, if analysis by statistically significant method, this center’s influence is not strong. To verify the third assumption, we have adopted a VAR model and impulse response analysis. B