House Market in Chinese Cities: Dynamic Modeling, In-Sample Fitting and Out-ofSample Forecasting

This paper attempts to contribute in several ways. Theoretically, it proposes simple models of house price dynamics and construction dynamics, all based on the maximization problems of forward-looking agents, which may carry independent interests. Simplified versions of the model implications are estimated with the data from four major cities in China. Both price and construction dynamics exhibit strong persistence in all cities. Significant heterogeneity across cities is found. Our models out-perform widely used alternatives in in-sample-fitting for all cities, although similar success is only limited to highly developed cities in out-of-sample forecasting. Policy implications and future research directions are also discussed.