欧债危机背景下各国股市间的波动溢出效应研究

发布时间：2018-03-17 02:31

本文选题：欧债危机　切入点：波动溢出效应　出处：《江西财经大学》2014年硕士论文　论文类型：学位论文

【摘要】：本文针对欧债危机前后全球11个国家股票市场间波动溢出效应的关系变化进行了深入研究,选取了英国、德国、法国、美国、中国、日本、意大利、西班牙、葡萄牙、希腊、爱尔兰、欧洲大盘的股票指数为研究对象。首先对各国股市的原始收益率数据进行BEKK-GARCH模型分析,但仅仅得到了各国溢出关系变化的粗略结论,不能看到不同时间尺度下的波动溢出效应,如高频数据间的波动溢出效应关系,长期趋势间的波动溢出关系变化。为了研究欧债危机背景下,各国股票市场间不同时间尺度下的波动溢出效应,本文对原始数据进行了多分辨率分解,得到各国股票指数的高频数据与长期趋势数据,然后对高频数据、长期趋势数据分别进行BEKK-GARCH模型溢出效应分析,从而得到不同时间尺度下各国股市间的波动溢出关系变化情况,并分析各国股市不同时间尺度下波动溢出效应的特点。目前常用的多分辨率分解方法是小波分解,但由于EMD方法具有比小波变换更强的局部表现能力,所以在处理非线性、非平稳信号时,是一种更为有效的方法。为了说明EMD分解对数据具有更强的表现能力,本文对EMD分解后的高频数据与小波分解后的高频数据进行对比,结果发现EMD分解后对数据的局部表现能力更强,而小波分解后局部表现能力不强且样本数量会减少一半,EMD分解后的长期趋势具有单调性,可以更清楚地刻画数据序列的长期走势,而小波分解后的低频序列仍无明显的规律。将原始收益率数据的波动溢出效应分析简称为第一层次分析,将不同时间尺度下的波动溢出效应分析简称为第二层次分析。通过两个层次的对比分析发现,两个层次的研究均得到了希腊是危机的主要输出国,葡萄牙是危机的主要传递国,美国与其他国家间的溢出效应最强。但仅通过第一层次的分析得出的结论较为粗糙,如通过第一层次得到的结论是“欧猪五国”与美国、中国、日本间的溢出效应整体上减弱了,但通过第二层次的分析得到的结果是“欧猪五国”高频序列与美国、中国、日本股市的高频序列溢出效应间呈现出不规律变化,而长期趋势间的溢出效应则呈现规律变化：希腊、意大利、西班牙与美国、中国、日本股市间的联动性较强,危机后仍呈现相互间的波动溢出效应,爱尔兰、葡萄牙向美国、中国、日本股市的溢出程度较弱,而更大程度地受美国、中国、日本股市向本国的波动溢出效应影响。另外,通过第二层次的多尺度溢出效应分析后发现,有一部分结论通过对原始收益率数据进行溢出效应分析并未得到,而因EMD分解对数据具有更强的表现能力,得出的结论更为细致,如中国股市的“脆弱性”,美国与英国、中国与德国、日本与英国、英国与德国股市间存在着长期趋势间的相互溢出效应。从11个国家间的溢出效应结果分析发现,一个国家的经济越发达,金融市场越稳定,在危机发生后,该国股票市场则更多地向其他国家产生溢出效应,而不是受其他国家单向溢出效应的影响。但由于我国仍处于发展中阶段,金融市场较为“脆弱”,在危机发生后,向其他国家的溢出效应由危机前的显著变为不显著,而受其他国家股市的溢出效应在危机后增强了。最后,论文针对我国金融市场的“脆弱性”提出了金融市场发展的相关建议。
[Abstract]:The paper further study the change before and after the debt crisis in 11 countries in the world stock market volatility spillover effect were selected, the UK, Germany, France, the United States, Japan, Italy, China, Spain, Portugal, Greece, Ireland, the European stock market index as the research object. Firstly, the original income from stock market rate data were analyzed by BEKK-GARCH regression model, but only a rough conclusion change of every spillover relationship, can not see the volatility spillover effect under different time scales, such as the volatility spillover effect between the high frequency data, change the long-term trend between the volatility spillover relations. In order to study the European debt crisis, the volatility spillover effect between the stock markets of all countries under different time scales, the multi-resolution decomposition of the original data, get high frequency data for different stock index and long-term trend data, then Based on the analysis of BEKK-GARCH model spillover effect of high-frequency data and long-term trend data, we get the volatility spillover relationship of stock markets in different time scales, and analyze the characteristics of volatility spillover effects of different stock markets in different time scales.
The current commonly used multiresolution decomposition method is wavelet decomposition, but because the EMD method is better than the wavelet transform local performance, so in dealing with nonlinear, non-stationary signal, is a more effective method. In order to illustrate the decomposition of EMD has stronger the ability of data, this paper compares the high frequency data of EMD the high-frequency data and after wavelet decomposition, the results showed that local expression ability of data after the decomposition of EMD, and the wavelet decomposition of local performance ability and the sample number will be reduced by half, the long-term trend after the decomposition of EMD is monotone, the long-term trend can clearly describe the data sequence, and the low frequency sequence after wavelet decomposition is still no obvious regularity.
The analysis will be referred to as the first level analysis of the volatility spillover effect of data rate of the original income, will the volatility spillover effect under different time scales analysis referred to as the second level. Through the comparative analysis of two levels of analysis, study two levels were obtained in Greece is a major exporter of crisis, Portugal is the main transfer in crisis the United States and other countries, the spillover effect is strongest. But only through the first level of analysis the conclusion is rough, such as the first level of the conclusion is "PIIGS" with the United States, Japan Chinese, spillover effect between the overall weakening, but the results obtained through the analysis of the second level is "Ou pig" high frequency sequence and high frequency sequence Chinese, America, Japan's stock market spillover effect between showing no regularity, and the spillover effect between the long-term trend has changed: Greek law La, Italy, Spain and the United States, Chinese, Japan stock market linkage is stronger, after the crisis is still showing volatility spillover effects between Ireland and Portugal to the United States, Chinese, Japan's stock market spillover is relatively weak, and to a greater extent by the United States, China, Japan stock market volatility spillover effect to its influence.
In addition, the multi-scale analysis of spillover effect of the second level after the discovery, some conclusions through the analysis of spillover effects has not been on the original return data, due to the decomposition of EMD has stronger data performance, the conclusion is more detailed, such as Chinese stock market's vulnerability, the United States and the United Kingdom, Chinese with Germany, Japan and Britain, the British and German stock market there are spillover effects between the long-term trend.
From the results of the spillover effect between the 11 countries of the analysis found that the more developed the economy of a country, the financial market is more stable, after the crisis, the stock market is more spillover to other countries, and is not affected by other countries. But the spillover effects in China is still in development stage, the financial market is "fragile", after the crisis, the backward spillover effects in other countries by significant change before the crisis was not significant, but by the spillover effects of the stock market in other countries increased after the crisis. Finally, according to China's financial market "vulnerability puts forward relevant suggestions on the development of financial market.".

【学位授予单位】：江西财经大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F831.51

【参考文献】