基于GARCH族模型的黄金现货价格波动研究

发布时间：2018-06-01 20:46

本文选题：国际黄金价格 + 美元指数　；参考：《天津财经大学》2013年硕士论文

【摘要】：黄金是国际公认的硬通货,它具有商品与货币的双重属性,是投资者资产保值增值的重要方式。最近十年,受欧洲主权债务危机和国际金融危机等因素的影响,黄金价格屡创新高出现了大幅波动,一直在高位震荡上行。越来越多的投资者将货币换成黄金套取增值,而黄金价格的剧烈震荡也让许多投资者遭受了巨大的损失。对黄金价格的波动性进行研究并从中找到金价波动的特点、对黄金价格未来的走势进行有效的预测,有着极其重要的理论和现实意义。文章首先详细介绍了国际上主要的黄金市场,深入分析了影响黄金价格波动的各个重要因素；然后选取合适的数据样本并进行数据预处理,通过对比分析建立预测模型并进行模型检验和预测效果的对比；最后总结全文并展望未来金价的走势。黄金价格波动预测的方法上多元回归模型使用的最多,但是该模型没有考虑残差的异方差性,效果并不理想。而时间序列模型则是根据被预测变量过去变化的规律性来建立模型,但是没有考虑外生变量的影响,对外界响应不敏感。所以,文章选取2002年1月2日至2013年2月22日的伦敦黄金交易所黄金的日定盘价作为样本,在模型上引入GARCH族模型来处理黄金价格序列中的条件异方差现象,在均值方程中引入对黄金价格影响最大的美元指数作为外生变量,最终建立起带有外生变量的EGARCH-X预测模型。结果显示,EGARCH-X模型不但可以有效地实现神经网络的黑盒子预测功能、而且大大提高了预测精度,可以直观地把握波动率变化关系,取得了很好的效果。文章结论表明,黄金价格的波动率存在着明显的ARCH效应,杠杆效应也很明显,并且波动率存在非常强的长记忆性。
[Abstract]:Gold is an internationally recognized hard currency, it has the dual attributes of commodity and currency, and is an important way to maintain and increase the value of investors' assets. In the last decade, gold prices have fluctuated sharply at record highs due to factors such as the European sovereign debt crisis and the international financial crisis. A growing number of investors are trading their currencies for gold, and many have suffered huge losses from sharp swings in the price of gold. It is of great theoretical and practical significance to study the volatility of gold price and find out the characteristics of gold price fluctuation and to predict the future trend of gold price effectively. In this paper, the main international gold markets are introduced in detail, and the important factors that affect the gold price fluctuation are analyzed in depth, and then the suitable data samples are selected and the data pretreatment is carried out. The prediction model is established through comparative analysis, and the model test and prediction effect are compared. Finally, the paper summarizes the full text and looks forward to the trend of gold price in the future. Multivariate regression model is widely used in forecasting gold price fluctuation, but the model does not take into account the heteroscedasticity of residual, so the effect is not satisfactory. The time series model is based on the regularity of the past changes of the predicted variables, but it does not take into account the influence of exogenous variables and is insensitive to the external response. Therefore, the paper selects the daily fixed price of gold from January 2, 2002 to February 22, 2013 as the sample, and introduces the GARCH family model to deal with the conditional heteroscedasticity phenomenon in the gold price sequence. The dollar index, which has the greatest influence on gold price, is introduced into the mean equation as an exogenous variable. Finally, the EGARCH-X prediction model with exogenous variables is established. The results show that the EGARCH-X model can not only effectively realize the black box prediction function of neural network, but also greatly improve the prediction accuracy. The conclusion shows that the volatility of gold price has the obvious ARCH effect, the leverage effect is also obvious, and the volatility has very strong long memory.
【学位授予单位】：天津财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F831.54;F224

【参考文献】