Time-Varying Mixture Model with Midas Technique for Measuring the Risk and Capturing the Contagion

Determining the distribution of financial assets and capturing macro and micro information have been the keys to financial risk management. In this paper, we propose a new model that simultaneously considers a mixture of distributions, stock market information, and macrofundamental factors to explore the impact of the US stock market and macroeconomic fundamentals on stock market risk in other economies. Specifically, the efficacy and feasibility of the model are first illustrated via simulations. We then apply the model to assess the contagion between the US economy and five other developed economies during four recent crises, and we forecast their VaRs. Our empirical results show that the risk of most test markets is negatively correlated with the macroeconomic fundamentals of the US and the US stock market. Furthermore, risk contagion from the US to the other tested stock markets is evident in the crisis years of 2002 and 2008. Meanwhile, the proposed model provides accurate forecasting of daily VaR. The findings emphasize that the new model is a flexible and reliable alternative for detecting and managing financial risks