分位数回归模型及其在黄金市场分析中的应用

发布时间：2018-01-13 05:08

本文关键词：分位数回归模型及其在黄金市场分析中的应用　出处：《合肥工业大学》2014年硕士论文　论文类型：学位论文

【摘要】：黄金兼具金融和商品两种属性，它是国家货币的储备金，是个人金融资产投资保值的工具。20世纪70年代以后，由于黄金价格的市场化，其价格发生了剧烈的变动，从70年代每盎司30美元一度涨到2011年的1900美元。2011年至今，黄金价格也经历了几次反复地涨跌，其价格的变动一直是各方关注的焦点。由于黄金自身的特殊性和重要性，国内外许多专家学者对黄金价格进行了实证和理论研究。近几年来，我国的黄金产量和黄金需求量都居于世界前列，因此研究我国黄金价格的影响因素具有重要的理论研究意义和实际应用价值。在分析总结国内外学者对影响黄金价格因素探究的基础上，，本文利用分位数回归方法探讨了原油价格、道琼斯指数、美元汇率、上证指数和利率对国内黄金价格的影响。实证结果表明道琼斯指数和美元汇率与黄金价格显著负相关，且影响比较稳定；上证指数和利率与黄金价格显著正相关，但利率的影响波动较大不稳定；而原油价格在该模型中对黄金价格的影响不稳定；利率和原油价格影响的不稳定表现在：当黄金价格较高或较低时，其影响相对较大，而在一般情况下，它们的影响相对较小。相对于最小二乘回归模型得到单一的结果来说，利用分位数回归得到的结果能够从历史数据中挖掘更多的信息，更有利于投资者了解黄金价格的波动因素，做出更好的决策。
[Abstract]:Gold has both financial and commodity attributes, it is the reserve of the national currency, is the tool of personal financial assets investment after 70s, due to the marketization of gold prices. Its price has changed dramatically, rising from $30 an ounce on 70s to $1, 900 in 2011, and gold has gone through several ups and downs since 2011. Because of the particularity and importance of gold itself, many experts and scholars at home and abroad have carried out empirical and theoretical research on the price of gold. China's gold production and demand are among the most important in the world, so it is of great theoretical significance and practical value to study the influencing factors of gold price in China. On the basis of analyzing and summarizing the factors influencing gold price, this paper discusses the crude oil price, the Dow Jones index and the dollar exchange rate by using the quantile regression method. The empirical results show that the Dow Jones Index and the US dollar exchange rate have a negative correlation with the gold price, and the impact is relatively stable; Shanghai stock index and interest rate are positively correlated with gold price, but the influence of interest rate is unstable. The influence of crude oil price on gold price is unstable in this model. The impact of interest rates and crude oil prices is unstable: when gold prices are higher or lower, the impact is relatively large, and in general. Compared with the least square regression model, the results obtained by quantile regression can extract more information from historical data. More conducive to investors to understand the volatility of gold prices, to make better decisions.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.54;O212.1

【参考文献】