基于GED—GARCH模型的中国原油价格波动特征研究

Study on the Characteristics of Chinese Crude Oil Price Volatility Based on GED-GARCH Models

本文采用中国大庆原油价格日平均交易数据,建立了基于GED分布的GARCH(1,1)、GARCH-M(1,1)和TGARCH(1,1)三个模型,描述了中国原油价格与国际接轨以来的波动特征。实证结果表明,与国际油价类似,中国原油价格的波动也存在显著的GARCH效应,但其波动冲击的半衰期要比国际油价短,为5天。而且,中国原油收益率受到预期风险的负向影响,表明中国原油市场并非完全市场化运作,当然这种负向影响程度较小,约为8%。另外,中国原油价格的波动存在显著的杠杆效应,相同幅度的油价下跌比油价上涨对未来油价的波动具有更大的影响,前者是后者的1.7倍左右。最后,基于GED分布的GARCH模型比基于正态分布的GARCH模型能够更好地描述中国原油价格的波动特征,并且具有较好的预测能力。

Using daily data of average crude oil price in Chinese Daqing oil market, this paper developed three kinds of GARCH model based on Generalized Error Distribution （ GED）, i. e, GARCH （ 1,1 ）, GARCH - M （ 1,1 ） and TGARCH（ 1,1 ）, so as to depict the volatility characteristics of Chinese crude oil price since its unification with the international oil markets. The empirical research result shows that, similar to the international oil market, there also exists significant GARCH effect in the price volatility of Chinese crude oil, and the half - life of its volatility shock is 5 days, shorter than that of the international crude oil. Further analysis indicates that the expected risk about return will have a weak but negative influence on the return with the extent about 8%, which implies that Chinese crude oil market has not been operated in the fashion of pure market economy. In addition, significant leverage effect can be found in the price volatility of Chinese crude oil market. Specifically, downward movements in the oil price will exert about 1.7 times larger impact on the following oil price volatility than that of upward movements with the same magnitude. Finally, these GARCH models based on GED can better describe the price volatility of Chinese crude oil compared with those based on the Normal distribution, and have great power to forecast the future returns.