Do economic policy uncertainty and its components predict China's housing returns?

In this paper we first studied the predictive power of economic policy uncertainty (EPU) by analyzing it in the traditional housing returns model, which included macroeconomic variables. Then we expanded the model to consider structure breaks and asymmetry and explored the optimal predictive model for housing returns. Last, based on the optimal model, we tested the predictive performance of specified EPU components. These analyses were employed on aggregate and different tier cities for the China housing market by using the flexible generalized least squares estimator to conduct in-sample and out-of-sample forecast evaluations. Empirical results revealed that our proposed model suggests EPU is superior to the traditional housing returns model in terms of in-sample and out-of-sample forecasts. We also found that extending the model to consider structure breaks and EPU asymmetry enhanced the prediction performance of the model. In addition, empirical results also showed that monetary policy uncertainty has the strongest predictability on housing returns in first-tier cities, while fiscal policy uncertainty has the strongest predictability on housing returns in the country, the second-tier, and third-tier cities.