On Volatility Spillover Effects And Risk Contagion Mechanisms In Real Estate Markets

With the acceleration of globalization of the world economy, the spatial interrelationships and volatility spillover effects among global financial markets are becoming increasingly strong and evident. Understanding the nature of the dependence structure and heterogeneity among global financial markets, especially global real eatate markets, hold important implications for financial experts, policy makers, and academic researchers. As such, what is the dependence structure among different markets, what drives the price comovement across markets and how to model volatility spillover effects and explore the risk contagion mechanism across markets are still hot and difficult for academic researchers.Based on the carding and induction of the recent literature on the applications of the classical spatial econometric models and multivariate GARCH models in financial markets in detail, this paper regards the price comovement across markets as a starting point, proposes two new approaches to combine the spatial econometric models and the multivariate GARCH models, discusses the structure of our proposed models, stationary conditions of parameters and maximum likelihood estimation method, and applies the proposed models to deeply analyze the volatility spillover effects and explore the risk contagion mechanism among real estate markets, stock markets and foreign exchange markets.First, the present study empirically investigates price comovement and volatility spillover effects among global real estate markets, stock markets, and foreign exchange markets using spatial DCC-GARCH models. The results show that the dynamic conditional correlations among real estate markets, stock markets, foreign exchange markets, and across these markets are time-varying. The volatility spillover effects and risk contagion exist among these markets. In addition, in terms of regions, the intensity of comovements among European countries'financial markets is stronger than the intensity of comovements among Asia-Pasifi and Latin American countires'financial markets.Second, based on ARMA (1,1)-GJR-AGARCH (1,1) models, this article examines the impacts of the recent global financial crisis 2007-2009 on conditional volatilities and dependence structure among real estate markets, stock markets, and foreign exchange markets. The results suggest that the conditional variances of all of the assets increase markedly during the global financial crisis period. In terms of the leverage effects, the significant positive coefficient of leverage effects both in real estate markets and stock markets indicates that the leverage effects exist in these markets. However, we do not find evidence that the leverage effects exist in foreign exchange markets.Third, this paper investigates the impacts of the price changes of US dollar index on real estate markets, stock markets, and foreign exchange markets in different countries. The results indicate that the price changes of US dollar index have significant influences on the dynamic conditional correlations among these markets. Particularly, the increase of US dollar index price will decrease the real estate market index prices in Asia-Pasific countries, but will promote the increase of the real estate market index prices among European and Latin American countires. In terms of stock markets, the increase of US dollar index price will promote the increase of the real estate market index prices among all the regional countries. However, it seems that the comovements between US dollar index and European and Latin American countries'stock markets are stronger than that between US dollar index and Asia-Pasific countries'stock markets. In terms of foreign exchange markets, the comovements between US dollar index and European Euros, Japanese Yen and British Pound are stronger than the comovements between US dollar index and Chinese Renminbi, HongKong Dollar, and Australian Dollar.Fourth, this article considers minimum-variance and hedged portfolios that are made up of two assets and evaluates the portfolios'efficiency using in-sample evaluation framework. The results show that both the minimum-variance and hedged portfolios have decreased the overall variance, the portfolio variances of these two portfolios are bigger in crisis period than in non-crisis periods. The proposed spatial DCC-GARCH model performs as well as those traditional models. Additionally, we find that the minimum-variance portfolio is more suitable for real estate market's assets and mixed assets that are made up of real estate market's assets, stock market's assets, and foreign exchange market's assets; whereas the hedged portfolio is more suitable for stock market's assets and foreign exchange market's assets.Finally, this paper generalizes a recently proposed dynamic spatial panel data model that accounts for multivariate asymmetrical GARCH components in disturbances. The specification criteria of the spatial weight matrices, stationary conditions of parameters, and maximum likelihood estimation method are also examined. Housing price index data on 10 cities for the period from January 2005 to December 2014 are used to investigate whether volatility spillover effects exist and to identify the factors that drive the price co-movement across regional housing markets in China. The empirical results show that stronger co-movement and volatility spillovers in housing returns appear among regions with geographic adjacencies or similarities of economic conditions. Fundamental factors such as population, income, and national macroeconomic situation are statistically significant determinants of housing prices in regional housing markets. Only the national macro-control policy on the housing market issued in May 2006 is found to have a significant impact on regional housing returns and fluctuations. The housing market segmentation between first-tier and second-tier cities has been more apparent since 2014. In addition, the presence of significantly positive leverage effects in regional housing markets suggests that investors often react more strongly to bad news than to good news when making property investment decisions.