黄金现货市场过度反应现象的实证研究及成因分析

发布时间：2018-04-21 21:13

本文选题：黄金现货市场 + 过度反应　；参考：《西南财经大学》2013年硕士论文

【摘要】：2006年到2008年国际黄金现货市场的成交量处于缓慢上涨阶段,最高的成交量最多也不过是10000多万手,但是2008年金融危机之后,黄金现货市场交易异常火爆,基本上成交量都超过了10000万手,甚至达到20000万手都不足为奇。而且2008年金融危机之后,每日的成交金额大约均在20万亿美元左右。这些数据说明,2008年金融危机之后,黄金现货市场成为热门投资地之一笔者在对黄金现货价格的不断追踪中发现,每当一个较为重磅的消息即将出来之前,黄金现货价格会上串下跳；在消息出来之后,立马趋势明朗,向同一个方向冲击；这期间的涨跌幅极大。很典型的几个例子,第一个,北京时间2012年2月29日晚,伯南克就美国经济未来经济决策讲话,因没有提到QE3,与之前市场的预期和相关推测相悖,黄金当晚跳水大跌,从1790.9美元/盎司跌倒到1687.0美元/盎司,跌幅接近100美元/盎司；第二个,2012年6月1日,美国公布5月份的非农数据,该数据显示5月美国非农就业增长6.9万,远低于市场预期的15万人；失业率为8.2%,高于市场预期的8.1%；因此QE3预期急剧升温,当日黄金价格迅速上涨,从消息公布时的价格1553.2美元/盎司,最高涨到1630.2美元/盎司,涨幅高达77美元/盎司。第三个,2012年8月31日到2012年9月13日,美国三大重磅消息：伯南克在世界央行大会的讲话、美国公布8月份非农数据、联邦公开市场委员会(The Federal Open Market Committee,简称FOMC)议息会议,推出开放式资产购买计划(即QE3),这三大消息全是利多黄金,跟之前的市场预期吻合,因此借势上涨,因此在黄金现货日线走势图中出现3大阳柱。但是这3大消息之后不是持续上涨而是经过一段盘整之后下跌。但是这三大消息出来之后,还是在短期内拉动黄金上涨100多美元。这样的例子不只三个,在笔者追踪期间,由于美国每月第一个周五公布非农数据,因此在几乎每个月的第一个星期五都有一次这样惊心动魄的画面。这种现象,让笔者质疑黄金现货市场可能存在过度反应。因为按照Fama(1970)提出的有效市场假说,这种波动幅度如此剧烈的情况应该是不会发生的。而且大量研究表明,金融市场上确实存在过度反应现象。那么黄金现货市场是否也存在呢? 本文以黄金现货市场的一些事实为切入点,引出过度反应现象可能存在。然后选取2006年3月14日到2013年3月1日的伦敦交易所黄金现货每日收盘价以及对应的成交量作为样本。为了避免出现“周日效应”、“元旦效应”等节假日效应,本文剔出了美国盘与欧洲盘的节假日数据,并以每周平均价作为最终样本。由于金融市场存在价格波动过度和交易过度异象,本文定义了收益率冲击和成交量冲击。并借鉴陈向明(2007)的过度反应检验模型,并结合黄金现货市场的实际情况,对其模型进行了改进,然后对黄金现货市场过度反应现象进行检验。过度反应模型的回归结果表明：利空消息时,交易量冲击和收益率冲击都会黄金市场造成过度反应现象,只是过度反应现象不同。其中收益率冲击造成的黄金现货市场的影响大于交易量冲击对黄金现货市场的影响。两种冲击对黄金现货价格的影响情况表现为：1)交易量冲击：首先,当期交易量冲击对黄金现货市场冲击微弱,这源于投资者相互之间并不知道对手交易量的多少,无法经过交易量冲击判断对黄金现货市场的影响；但是前一期交易量冲击对黄金现货市场冲击明显,原因是前一期的交易量冲击是已知的。其次,滞后的第18周利空消息造成对黄金现货市场的反应过度现象在滞后的第1周的时候得到修正,而不是当期等到修正,且引起的过度反应程度没有收益冲击大。2)收益率冲击：首先,当期收益率冲击对黄金现货市场冲击较大,这可能源于当期信息被投资者知晓后对黄金现货价格走势做出判断,从而使得价格承载了不少当期信息,进而反映在收益率中,影响当期收益率冲击。其次,滞后的第11周的利空消息造成对黄金现货市场的反应过度现象在当期得到修正,而由于当期的相互影响系数绝对值较大,故而修正的幅度较大,反应过度的程度较大。但是利好消息时,无论是交易量冲击还是收益率冲击,都准确的反映到市场中,没有出现反应不足的现象。黄金现货市场存在过度反应现象说明可以通过这些现象的特征对未来价格进行预测,就与Fama的弱式有效市场的特征相背离,因此,黄金现货市场是非有效市场。消息方面,本文认为,黄金现货价格实际上对未来世界经济形势的预期,因此能对未来世界经济形势做出推测的消息才是与黄金市场有关的消息；对黄金现货市场来说,利空消息是对世界经济有正面积极影响的消息,而利好消息是对世界经济有负面影响的消息。对过度反应现象解释的原因方面,本文认为,交易制度,特别是保证金制度,以及非理性偏好和心理偏差都是导致过度反应的原因。对投资者建议方面,本文认为可以根据本文的结论,投资者可在消息搜集和风险控制方面,做出相应对策。例如消息搜集方面,不要局限于本期的消息,也要注意前面的消息,便于分析世界经济环境状况；同时对世界经济环境有影响的消息不要露掉。风险控制方面,注意杠杆的运用,止损的设置,以及自己非理性行为的控制。本文在实证中,主要使用了自相关和偏自相关图、自相关模型、ARMA模型以及ARCH模型等计量方法。本文的创新点在于：一、研究领域的创新；黄金现货市场相对于股票市场和期货市场,研究热度不高；而且大量文献只针对证券市场、期货市场的过度反应现象研究,对黄金现货市场的过度反应研究几乎无人涉足。二、使用方法的创新；本文虽然借鉴了陈向明(2007)的模型,但是在他的基础上,根据黄金现货市场的特点进行了调整。三、解释过度反应现象的创新；本文不单从行为金融学理论研究成果去解释过度反应现象,也从黄金市场的交易制度、信息对称性和信息传导情况等方面对过度反应进行了解释。而且并没有利用行为金融理论有关过度反应解释的模型,而是利用黄金现货交易量影响因素模型和黄金现货价格影响因素模型。检验出这两个模型存在异方差现象,并结合与金融市场相关的心理偏差去解释过度反应现象。本文的不足之处在于：一、本文使用的过度反应检验的方法上存在一些问题,例如黄金现货收益率受到前1期和前12期的影响,但是为了模型的简单化,选择了前一期与交易量冲击(或收益率冲击)的交互影响来作为核心检验项,而不是前12期或者前1期和前12期与交易量冲击(或收益率冲击)的交互影响来作为核心检验项；这样做的原因是前1期对黄金现货收益率的影响比前12对他的影响大。二、本文在解释利好和利空消息的时候,只是将这些对黄金现货市场的影响消息大致归类,并且分析浅尝辄止；没有详细的解释这些消息在什么情况对黄金现货市场有怎样的影响。三、在对过度反应成因分析的时候,由于本人能力有限,只是文字描述以及结合现象的描述,并没有用数据和实证模型去分析。本文章节安排：第一章绪论,主要是引出过度反应现象,简要介绍过度反应与黄金市场的研究综述,还介绍了本文目的、意义、研究思路和研究方法、创新点与不足,以及章节安排；第二章相关理论回顾,主要是介绍市场有效假说的演变与行为金融的发展,以及过度反应现象的理论研究与成因解释；第三章黄金现货市场过度反应现象的实证研究,通过数据分析检验国际黄金现货市场存在反应过度现象及过度反应现象特征,并且通过对黄金市场信息引起的价格变化现象的简要说明,对利好与利空消息进行归类；第四章分析国际黄金现货市场过度反应现象的成因,利用行为金融相关理论、黄金市场交易制度、信息对称性与传导情况等分析过度反应现象的成因；第六章主要是总结全文,并为黄金现货投资者在搜集信息与分析黄金现货价格走势,以及风险控制方面提出建议。
[Abstract]:From 2006 to 2008, the volume of the international gold spot market was in a slow rise, the highest volume was at most about 100000000 hands, but after the 2008 financial crisis, the gold spot market traded exceptionally well, basically over 100 million hands, and it was not surprising to reach 200 million. And 2008 financial After the crisis, the daily turnover was around $20 trillion. These figures show that after the 2008 financial crisis, the gold spot market became one of the most popular investments.
In the track of gold spot prices, I found that the spot price of gold will jump every time a heavier piece of news is coming out, and after the news comes out, the trend is clear and hit the same direction.
One of the typical examples, the first, in February 29, 2012, Beijing time, Bernanke spoke about the future economic decisions of the US economy. As it did not mention QE3, contrary to previous market expectations and related speculation, gold fell sharply, falling from $1790.9 / ounce to $1687 / oz, falling close to $100 / ounce; second, 2. In June 1st 012, the United States announced the non farm data of May, which showed that US non-agricultural employment increased by 69 thousand in May, far below the expected 150 thousand of the market; the unemployment rate was 8.2%, higher than the market expected 8.1%; so QE3 was expected to rise sharply, and the price of gold rose rapidly on the same day, up to 1 from the announcement price of 1553.2 US dollars / ounces, up to 1. 630.2 dollars / ounces, up to $77 / ounce, third, from August 31, 2012 to September 13, 2012, the three big U. S. news: Bernanke's speech at the World Bank Conference, the US announcement of non farm data in August, the The Federal Open Market Committee (FOMC), and the introduction of open assets. The purchase plan (QE3), the three big news of all the gold, coincided with the previous market expectations, so the loan rose, so the 3 big sun columns appeared in the gold spot chart. But the 3 news did not continue to rise but fell after a period of consolidation. But after the three news, it was still in a short period of time. Gold is up more than 100 dollars.
There are more than three examples of this. In the course of my tracking, as the United States published non-agricultural data on the first Friday on the first Friday, there was such a thrilling picture on the first Friday of almost every month. This phenomenon asked the author to question the possibility of overreaction in the gold spot market. In accordance with the effective market proposed by Fama (1970) The field hypothesis, such an intense volatility, should not happen. And a lot of research shows that there is a real overreaction in the financial market. Do the spot market for gold exist too?
This paper, taking some facts of the gold spot market as a breakthrough point, leads to the possibility of overreaction, and then selects the daily closing price and the corresponding volume of the gold spot on the London exchange from March 14, 2006 to March 1, 2013 as a sample. In order to avoid the "Sunday effect", "the new year's day effect" and other holiday effects, The article extracts the data of us and European holidays, and takes the weekly average price as the final sample.
Because of the excessive price volatility and excessive trade anomalies in the financial market, this paper defines the impact of yield and the impact of turnover. And we use Chen Xiangming's (2007) overreaction test model and combine the actual situation of the gold spot market to improve the model, and then test the overreaction phenomenon of the gold spot market. The regression results of the degree response model show that both the trading volume impact and the yield impact will cause overreaction in the gold market, but the overreaction is different, in which the impact of the yield impact on the gold spot market is greater than the impact of the transaction impact on the gold spot market.
The impact of the two shocks on the spot price of gold shows as follows: 1) the impact of trading volume: first, the impact of the current trading volume impact on the gold spot market is weak. This is due to the fact that investors do not know how much of the opponent's transaction between each other, and can not judge the impact on the gold spot market by trading volume impact; but the previous volume of trading volume. The impact on the gold spot market is obvious, because the volume shocks of the previous period are known. Secondly, the lagging eighteenth week profit message caused the overreaction to the gold spot market in the first week lag, not the current revision, and the degree of overreaction caused by no return to the big.2). The impact of interest rate: first, the impact of the current rate impact on the gold spot market has a great impact, which can be used to judge the gold spot price trend after the current information is known by the investors, thus making the price bearing a lot of current information, and then reflecting the impact of the current rate of return on the rate of return. Secondly, the lagging eleventh weeks of profit. The phenomenon of overreaction to the gold spot market is amended in the current period, and the absolute value of the mutual influence coefficient of the current period is larger, so the extent of the correction is larger and the degree of overreaction is greater. However, when the good news, whether it is the impact of the volume of transaction or the impact of the rate of return, it is accurately reflected in the market. A phenomenon of insufficient reaction.
The phenomenon of overreaction in the gold spot market shows that the future price can be predicted by the characteristics of these phenomena, and it deviates from the characteristics of the weak effective market of Fama. Therefore, the gold spot market is a non effective market.
In the news, this article believes that the gold spot price is actually expected in the future world economic situation, so the news that can speculate on the future world economic situation is the news related to the gold market; for the gold spot market, the sharp news is the news that the world economy has a positive impact on the world economy, and the good news is the world The news that the economy has a negative impact on the economy.
In terms of the reasons for the explanation of overreaction, this paper argues that the trading system, especially the margin system, and the irrational preference and psychological bias are the causes of overreaction.
According to the conclusion of this paper, investors can make corresponding countermeasures in terms of information collection and risk control. For example, information collection, not limited to the news of this period, should also pay attention to the previous news, to facilitate the analysis of the world economic environment; at the same time, the influence of the world economic environment will be eliminated. Do not lose interest. Risk control, pay attention to the use of leverage, stop loss settings, as well as their irrational behavior control.
In this paper, we mainly use the methods of autocorrelation and partial autocorrelation, autocorrelation model, ARMA model and ARCH model.
The innovation of this article lies in: first, innovation in the field of research; the gold spot market is not very hot on the stock market and the futures market; and a large number of literature only focuses on the overreaction of the stock market, the futures market, and the overreaction of the gold spot market is almost unrelated. Two, the innovation of the method of use. This article, on the basis of Chen Xiangming (2007) model, has been adjusted on the basis of the characteristics of the gold spot market. Three, to explain the innovation of overreaction; this article is not only from the research results of behavioral finance to explain the overreaction, but also from the trading system of the gold market, information symmetry and information. There is no use of the model of behavioral finance theory on Overreaction explanation, but the model of the impact factor model of gold spot trading volume and the influence factor model of gold spot price. Test the difference between the two models and the relationship with the financial market. The reason is to explain the phenomenon of overreaction.
The shortcomings of this paper are: first, there are some problems in the method of overreaction test used in this paper, such as the impact of the gold spot yield on the first 1 and the first 12 periods, but for the simplification of the model, the interaction of the previous period and the transaction volume impact (or the rate of return impact) is chosen as the core test, not the first 1. The interaction of the 2 or the first 1 and the first 12 periods with the trading volume impact (or the rate of return impact) is used as the core test; the reason for this is that the first 1 has a greater impact on the gold spot rate than the previous 12. Two, this article, when explaining good and good news, is just the news of the impact on the gold spot market. There is no detailed explanation of what effects these messages have on the gold spot market. Three, in the analysis of the causes of overreaction, because of my limited ability, only the description of the text and the description of the combination phenomenon, it is not analyzed with data and empirical models.
This chapter is arranged as follows: the first chapter is the introduction, mainly introducing overreaction, briefly introducing the overreaction and the gold market research, and introducing the purpose, significance, research ideas and methods, innovation and insufficiency, and chapter arrangement; the second chapter is a review of the relevant theory of the market, mainly introducing the evolution of the effective hypothesis of the market. The development of behavioral finance and the theoretical and genetic explanation of the phenomenon of overreaction; the third chapter is the empirical study of overreaction in the gold spot market. Through data analysis, the phenomenon of overreaction and overreaction in the international gold spot market is tested and the price changes caused by the gold market information are shown. The fourth chapter analyzes the causes of overreaction in the international gold spot market, analyzes the causes of overreaction by behavioral finance related theory, gold market trading system, information symmetry and transmission, and the sixth chapter is mainly to sum up the full text and to gold spot. Investors are making suggestions on gathering information and analyzing gold spot price movements and risk control.

【学位授予单位】：西南财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.54

【参考文献】