| In the process of global economic integration, the domestic enterprise's production and operating activities increasingly influenced by the price of international market raw material and commodity, especially in the 08 years after the outbreak of the financial crisis in domestic enterprises are facing tremendous commodity price risk. Especially after the outbreak of the international financial crisis in 2008, domestic enterprises are facing a huge risk of commodity price fluctuations. The companies are looking for ways to disperse the price risk. The market price of raw materials and products determine the enterprise's survival directly, so to avoid price risk is the prerequisite for the survival and development of domestic enterprises. The derivatives markets provide an effective solution to this program; an important function of futures market is locking the commodity prices and avoiding price risk through the futures hedge. The core of this study is using the function of hedge of futures contracts to spread the risk of price volatility. The article aims to compare among the different hedging models, find out the optimal hedge ratio and increase the effect of futures hedging.Current research of hedging strategy is based on GARCH model, but this model has a big problem, this model has serious shortcomings in characterize randomness. In this paper, we adopt multivariate stochastic volatility model (MSV model) to study hedge, multiple stochastic volatility model is a class of heteroscedastic models, and it introduced stochastic process into the model. Compared with GARCH, it will be better to reflect than the randomness of financial markets. It is the optimal models to depict financial market volatility.In this paper, firstly, we build minimum variance hedge model with CC-MSV model to study the optimal hedge ratio. Then based on CC-MSV model, we introduce dynamic correlation model and multivariate t-distribution. This new model is named DC-t-MSV. The model has considered the peak and fat-tail characteristics of financial data as well as the local correlation of the correlation coefficient between the spot and futures'logarithmic returns, which is more in line with the characteristics of financial time series. Finally, because of the vulnerability of financial market, the develop of financial market will suffered from unexpected events ,such as economic crises, natural disasters, changes in external economic policy and so on. So based on DC-t-MSV model, the article introduce Markov regime switching (regime -switching) to build the RSDC-t-MSV minimum variance hedge model to represent the sudden impact on the futures market. RSDC-t-MSV model has considered both the internal features of the financial data and external contingencies. It is consistent with the actual situation of financial markets.In the empirical part, the article uses logarithmic return of gold's spot and future price from January 10, 2008 to June 30, 2011 to do empirical research. The data are used for CC-MSV model, DC-t-MSV model and RSDC-t-MSV model to study on hedging; it also uses hedging measure formula to compare the hedging effect based on the three different models. During the empirical research, we begin with the use of Eviews5.0 to analyze the statistical characteristics of data, and then use winbugs to compute the model. In the part of Parameter estimation, the article applies slice sampling techniques of MCMC methods. This technology can be more effective compared with the Gibbs sampling.The empirical analysis of gold futures hedging efficiency has come to the desirable conclusion. Three models have obtained the dynamic optimal hedge ratio. All of them have achieved a satisfactory effect of hedging. However, comparative analysis shows that the optimal hedge ratio calculated by the RSDC-t-MSV model not only can capture the market volatility, but also ensure the stability of the hedge rate. It has great advantage compared with DC-t-MSV model and CC-MSV model. From the side of hedge effort, RSDC-t-MSV model can achieve a greater degree of reduction of the portfolio investment risk. It can do well in hedge than CC-MSV model and DC-t-MSV model. Therefore, we can consider that the use of RSDC-t-MSV model can do better in the area of hedge. The empirical analysis is fully consistent with the theoretical analysis. |
