住房市场模型研究——基于非稳定时间序列的FM-OLS估计

Housing Market Model Regression——FM Estimation on Nonstationary Time Series

通过对住房市场进行系统分析,提取影响房价和销售额的主要变量:房地产开发国内贷款、已开发的房产当前的存量、住宅施工面积、金融机构中长期贷款、储蓄存款余额和宏观经济环境变量GDP,建立住房市场房价与销售额回归模型。并通过FM-OLS方法对非平稳时间序列回归模型进行估计和统计推断,解决了传统OLS方法无法对非平稳时间序列回归进行统计推断问题,避免了对数据差分平稳化处理导致的信息丢失及回归系数不易解释的问题。实证结果表明房价和销售额受宏观经济环境GDP影响显著,房屋存量在方程中的表现揭示了中国房地产市场房地产开发商囤积房源和减少供给量导致住房市场价格持续上涨的现象。最后,通过与ARIMA模型拟合效果进行比较,证明了模型的拟合效果更优,更能体现市场的真实情况。

In this paper, through analysis of the housing market, the residential housing market system abstractly includes banks, house producers, and house buyers; each interacts with the others through transactions in the macroeconomic environment. We build up the Housing sales and house price regression model by extracting the main variables： the domestic real estate development loans, the stock of the current real estate, residential construction area, financial institutions and long - term loans, savings deposits and macroeconomie environment variable GDP. FM - OLS method estimation and statistical inference to solve the problem that the traditional OLS approach can not give valid statistical inference for non - stationary time series regression. Empirical results show that prices are subject to the impact of the macroe- conomic environment significantly, housing stock in the performance of the equation reveals the real estate market real estate developers in China hoarding housing to reduce the supply, resulting in the housing market prices continue to rise. With the comparison with ARIMA model, this paper proves that our model fits better.