Dependence structure and risk spillovers between real estate and stock markets: An application of VMD based time-varying copula approach

Abstract: In this thesis, we combine copulas with the variational mode decomposition (VMD) method to explore the dependence structure between real estate and stock market in three countries, namely China, U.S. and Australia. We explore the static and dynamic symmetric and asymmetric copulas, and investigate the time-varying dependence structures in the short-term and long-term time horizons. Our empirical results provide strong evidence of tail dependence between the stock market and real estate market in all countries. Furthermore, lower tail dependence is generally higher than upper dependence in all time horizons. We then quantify risk measures, namely value at risk (VaR), conditional VaR (CoVaR) and the delta CoVaR ( CoVaR) to analyze both upside and downside risk spillovers at different time horizons for each country. We find significant bidirectional risk spillovers between the stock market and the real estate market. The risk spillover effect between real estate-stock pair is the strongest in U.S and the weakest in Australia. In addition, we observe that downside risk spillovers are significantly stronger than the upside spillovers, and the systemic risk contribution of real estate to stock markets is larger than that of its opposite direction in all three countries.