Price Volatility Risk Assessment And Comparison Of China's Commodity Futures

Financial market risk is earnings deviate from the expected future earnings due to asset price fluctuations. After years of development, China's commodity futures market improved gradually, but the huge market price volatility in recent years resulted in huge losses to the futures market participants. So it is significant to estimate price volatility risk in commodity futures market.As the bulky capacity of Shanghai copper futures market, it attracts a large number of financial institutions including in investment funds. While as the most important industrial products, commodity price is impacted by the world economic development by large, therefore, a large number of commodity production and consumption enterprises involved in futures hedging operations to control price volatility risk. Shanghai copper futures market has a long history, and accumulated a large amount of historical transaction data; Meanwhile, trading margin for each hand of copper futures is relatively high which reduces the speculative risk of the market. Therefore, to analysis VaR of copper futures market can reflect other commodity futures market volatility risk.After years of development, scholars have developed a variety of risk measurement model. With which use the VaR method to estimate market volatility risk is popular in recent years. Unlike the previous risk measurement method, VaR method can give an exact value of loss in predetermined confidence level. This simple and intuitive method of risk measurement has been widely welcomed by financial institutions. In this paper, by comparing the GARCH-N, GARCH-T, GARCH-GED, POT-GPD model we try to estimate the VaR of copper and test the accuracy of various models by use Kupiec test.Following conclusions are obtained in this paper:China's commodity futures price return series tend to Heavy-Tailed and Lep-tokurtosis.Shanghai copper futures return series do not exist leverage. Which means the impact of good news and bad news on the return volatility of copper is no different. In the low confidence level, VaR based on student T distribution, GED distribution, and extreme value theory fit results are more satisfactory; but in high confidence level, GARCH_T, GARCH_GED models and extreme value theory under the VaR can be more accurate fitting return loss. The GARCH model under normal distribution, whether in high or low confidence level are not well cover the actual loss.