The Analysis And Forecast Of Price Of Guangzhou Real Estate

The real estate is the basic of national economy, and is closely related with people's lifequality, and is linked with many industries of the national economy. It plays a decisive role inthe modern economy society. The commercialization of housing is not long in China's history,but in recent years, the trend of the rapid rising in housing prices has tremendous impact oncurrent economic and society. In order to control prices, the government has adopted a seriesof policy. It's a very meaningful topic that we use theoretical and empirical analysis on thechange tendency of price.This article takes the Guangzhou city(contain ten districts and Conghua city andZengcheng city) and ten districts of commercial housing sales prices(sales divided by thesales area)as the research object. We relatively complete collected the related data of thehousing prices over the years, and analysed the price changes of Guangzhou city. The work ofthis paper are as follows. First, we analysed the weak-form efficiency of Guangzhou realestate market. Using a variety of statistical testing method to the housing price, the mainlymethods are serial correlation test and runs test. We should check that the market ofGuangzhou city and ten districts reject or not reject the weak-form efficiency. Our empiricalresults are that except Guangzhou city and Panyu, the other nine districts could not meet theweak-form efficiency. The conclusion may be the basic of that we analysed the forecast ofprice change. Second, we set up the house price changes model to the market that rejected theweak-form efficiency. First of all, we used the house price data of the nine district to do thestationarity test. We mainly used unit root test to the nine series of the logarithm of houseprices. And the results show that they are first order difference stable. Then according to theB-J methodology we established ARIMA model to the difference series of the logarithm ofhouse price, and used the residual correlation to test the model fitting effect. Last we usedthe ARIMA model to forecast, according to the real data of2011.9-2012.2, we compute theproportional error and find that ARIMA model can forecast very good in short-term. Next, wealso established the state space model to predict the house price changes. The results showsthat the state space model of the prediction effect is more excellent than the ARIMA model.