【作者】 高艳；

【导师】 陈守东；

【作者基本信息】 吉林大学 ， 数量经济学， 2014， 博士

【摘要】 汇率是各国进行贸易、资本跨国流动的前提，是连接各国经济往来的桥梁，因此汇率波动会对各国的经济产生一定的影响。汇率波动对经济波动及国际金融风险都具有良好的时效性和敏感性，收益率的变化会影响汇率的风险暴露。研究汇率的波动特征能更好地把握汇率变化的规律，从而化解汇率波动对经济的冲击。反过来，经济的变化也会对汇率收益率波动造成冲击，影响汇率收益率的波动特征。我国采用有管理的浮动汇率制度，并逐渐放开汇率的波动幅度，以使人民币汇率波动更加灵活，保证我国对外贸易和资本流动的稳定性。因此在研究均衡汇率和汇率传导机制的理论背景下对汇率的波动特征及计量方法进行深入的分析，可以对我国汇率的市场化深入改革提供理论支持。本文在人民币汇率均衡决定理论及国内外学者对汇率波动问题研究的基础上，对汇率波动的分布特征、人民币汇率波动对中美宏观经济的影响、人民币兑美元、欧元等汇率的协同波动溢出效应、波动协同持续特征的计量分析方法进行了系统深入的研究。首先，本文深入研究了汇率波动对经济的传导机制，包括汇率对价格水平、贸易条件、国外直接投资、就业、利率等的传导效应，其中，汇率对价格水平的传导包括直接传递效应和间接传递效应，汇率的变动在进出口时点上将直接影响进口或出口商品的价格，随后，价格波动将通过生产、贸易等途径传递到两国国内商品价格中，进而影响整个国家价格水平的变动，但是汇率的价格传递效应是不完全的。汇率对贸易的传导主要包括汇率水平变动通过价格机制影响出口国家的国际竞争力及波动带来的风险通过影响出口厂商贸易决策而影响进出口。汇率对国外直接投资传导的理论主要有“相对生产成本效益”理论和“相对财富效应”理论。汇率是国家之间贸易的纽带，因此汇率不仅会对国内外经济产生影响，也会受到国内外宏观经济的影响，本文就基于“模型学习”方法，利用Taylor规则实证分析人民币汇率与中美经济之间的动态相关性，从而在不同的汇率波动状态下，刻画人民币汇率受两国宏观经济的影响，实证结果表明，汇率的不同波动时期，宏观经济变量和汇率之间的动态关系不同。汇改初期汇率波动较大，汇率主要受中美产出缺口与中美利率差的影响；而金融危机期间我国采取软盯住美元政策，汇率波动较小，影响汇率变动的因素包括中、美两国的产出缺口，美国的通货膨胀率和中美利率差。第二次汇改重新启动后，所有引入模型的宏观经济变量均对汇率产生动态性影响，说明与第一次汇改相比，我国与世界经济结构已发生变化，汇率市场逐渐成熟，与国际接轨。其次，在对多元GARCH模型进行深入研究的基础上，重点研究了如何将一般基于正态分布的多元GARCH模型扩展为基于广义误差分布的多元的分布特征的GARCH模型。汇率的波动特征主要包括高峰厚尾特征、波动的集聚性特征、波动的长记忆性和持续性、杠杆效应、均值回复效应和波动溢出效应等，同时刻画各种汇率波动特征的基于正态分布、学生t分布、广义误差分布的GARCH模型，SV模型及区制转移模型，并运用二元GED-GARCH模型对汇率与利率之间的波动溢出效应进行了研究，通过自适应绝对偏差和自适应均方误差的平方根两种标准对基于正态分布的二元GARCH和二元GED-GARCH模型进行了比较，发现后者预测效果更好，并得出结论：在金融危机前利率与汇率之间存在着汇率到利率的溢出效应，而在金融危机之后，利率与汇率具有双向的波动溢出效应。第三，本文对人民币汇率的波动溢出效应和协同波动溢出效应进行了系统的研究。一方面通过运用VMEM方法研究了中、日、韩三国实际有效汇率的波动溢出效应，发现中国和日本的汇率波动既受自身过去波动、收益率的影响，也存在显著的杠杆效应；韩国汇率波动受前期波动与收益率的影响，但是不具有显著的杠杆效应。中日、日韩的汇率市场之间存在着双向的波动溢出效应，中国与韩国汇率市场存在单向波动溢出效应。同时，三个汇率市场的波动之间不存在交互的杠杆效应。另一方面将独立成分分析方法引入GARCH模型，度量了人民币兑美元、欧元等多个双边汇率的协同波动溢出效应。通过计量分析，得出结论：将人民币兑港元、日元、欧元及英镑汇率波动作为一个系统度量其对美元汇率的协同波动溢出效应时，只表现出低波动下港元对美元的波动溢出效应；通过投资美元、欧元及英镑汇率可以规避港元汇率的高波动风险；欧元及英镑容易受到其他汇率市场的波动影响；日本经济发展的特殊性，导致其他汇率市场对日元汇率不存在显著的波动溢出效应。最后，本文对人民币兑美元、欧元等汇率的波动持续性及协同持续特征进行了深入的研究。其中，人民币兑美元汇率存在着均值长记忆性及方差长记忆性的双记忆性特征，然后又对人民币对欧元、英镑、澳元、加元及日元的波动持续性及协同持续性进行了计量分析，发现人民币对非美元汇率的波动持续性均强于人民币兑美元汇率的波动持续性，主要原因是人民币对其他汇率是通过人民币兑美元汇率，与美元兑其他货币汇率套算得到的，持续性是两者的综合，所以表现出更强的持续性，通过借鉴协整理论的思想对其他汇率的波动协同持续性进行了计量分析，发现人民币对欧元、日元、英镑等汇率存在着明显的协同持续特征，即单个汇率市场波动是持续的，但是多个汇率市场的综合效表现出了非持续性，或较弱的持续性。人民币兑欧元汇率与人民币兑日元汇率之间的协同持续性最强。该分析结果对汇率的投资组合决策有很重要的意义。更多还原

【Abstract】 Exchange rate is the precondition for international trading and capital flow betweendifferent countries, it also is the bridge linking economic exchange for each country, so exchangevolatility will have certain affect to each county’s economy. Exchange volatility has bettertimeliness and sensitivity for economic volatility and international financial risk, change in fieldwill lead to risk exposure. Study of volatility features for exchange rate can lets us to know thelaw of the change for exchange rate, so that the shock from exchange rate volatility to economycan be defused. In reverse, change in economy can also cause shock to field volatility ofexchange rate, and affect the volatility features of exchange rate. Managed floating exchangesystem is adopted in our country, and the fluctuation range is released to bigger, so that the RMBexchange rate volatility become more flexible, which guarantee the stability of foreign trade andcapital flows. Therefore, on the basis of studying equilibrium exchange rate and exchange ratetransmission mechanism, analysis of the volatility features and quantitative methods forexchange rate deeply can be used as the theory support supplied for marketization reform deeplyof China’s exchange rate.In this article, on the basis of exchange rate decision theory and studiesof scholars at home and abroad for exchange rate volatility problems, we will analyze thedistribution features of exchange volatility, the effect of RMB exchange rate to china and USeconomies, the common volatility spillover effect and common volatility persistence betweenRMB/US, RMB/GBP and other exchange rates.First, transmission mechanism from exchange rate volatility to economy will be analyzeddeeply, which contains exchange rate to price level, trading condition, FDI, employment, interestrate and so on. In which, the transmission to price level include direct and indirect transmissioneffects. Change in exchange rate will affect the price of import and export commodities directly,then the volatility of price will transmit to domestic commodities prices by production and trade,so that affect change of the whole countries’ price level, but the price transmission is notcomplete. Transmission from exchange rate to trading is mainly contains two aspects:1) exchange rate change will affect the international competitiveness through price mechanism;2)the risk brought by volatility will affect the export and import by affecting decision of the exportmanufacturer.Theory of transmission from exchange rate to FDI mainly has “Relative cost-effectivenessand “relative treasure effect”. Exchange rate is the link of trade between countries, so exchangeis not only has effect to domestic economy, it also affect other country’s economy, Dynamiccorrelation between exchange rate and fundamental economic variables is always focused on.Inthis article, based on “Model learning method” to describe the problems. Empirical study turnsout: the dynamic relationship between RMB/US exchange rate and fundamental economicvariables vary in different volatility horizon. At the beginning of the first exchange rate reform,the exchange rate is mainly affected by output gas and interest rate differential. During thefinancial crisis, our country peg dollar softly with low volatility, the influence factors are morethan before. After the second exchange rate reform, the whole world’s and our country’seconomic structure has changed, all variables in model affect the variation of the exchange rate.Second, on the basis of studying multiple GARCH model deeply, we focus on how toextend the multiple GARCH model based on Normal distribution to multiple GARCH based onGED, Volatility features of exchange rate mainly contains high kurtosis fat tail, volatilitycluster, long memory and persistence, leverage effect, mean reversion and volatility spill overeffect, and so on. At the same time, we will describe the volatility models, such as GARCHmodel based on Normal, student t distribution and GED, SV model and Markov Regimeswitching Models. We use binary N-GARCH and GED-GARCH models to analyze the SpilloverEffects between Exchange Rate and Interest Rate before and after financial crisis separately, andthen we evaluate the two models by Adaptive mean absolute deviation and adaptive root of meansquare error criterion. As a result, we think the forecasting effect of binary GED-GARCH isbetter, and we conclude from this model that there is no Spillover Effects between ExchangeRate and Interest Rate before financial crisis, but after financial crisis, there are two-waySpillover Effects between them.Third, volatility spillover effect and common volatility spillover for RMB exchange ratesare studied. On on side, Vector Multiplicative Error Model （VMEM） is used to study theexchange rate volatility spillover effects between China, Japan and South Korea. Empiricalanalysis results indicate: the real effective exchange rates for China, Japan is not only affected bytheir own fluctuation and return in last term, but also have significant leverage. For volatility ofSouth Korea, it is affected by its fluctuation and return in last term, there is no significantleverage, there are mutual volatility spillover effects between China and Japan, Japan and South Korea exchange rate markets. There is no volatility spillover effect from China to South Korea，but has volatility spillover effect from South Korea to China. At last, there is no mutual leverageeffect between exchange rate markets. On the other side, I measure common volatility spilloverfrom some exchange rate markets to one by introducing ICA method into GARCH models.Empirical evidences show that,1) if volatilities of HKD, JPY, EUR and GBP are taken as asystem to measure the common volatility spillover to USD, there is volatility spillover fromHKD to USD only.2) We can avoid high volatility risk by investing USD, EUR and GBP.3)EUR and GBP tend to be influenced by other exchange rate markets.4) There are no volatilityspillovers from other exchange rate markets to JPY/CNY because of the particularity of Japaneseeconomy.At last, the volatility persistence and common persistence for RMB exchange rates has beenstudied deeply. By analysis, we find the RMB/US exchange has both mean and variance longmemory, then RMB to EUR, GBP, AUD, CAD and JPY exchange rates’ volatility persistenceand common persistence have been economically analysis, we find the persistence of other nonRMB/US exchange rate is stronger than RMB/US exchange rate, I think the main reason isExchange rate of RMB to other money is cross-rate through RMB/US, persistence degree is theintegration of two exchange rates, so it show stronger persistence. Reference to theCo-integration theory, I analyze the common persistence for other exchange rates except forRMB/US by economical method, then I find, there are common persistence between otherexchange rate, namely, each exchange rate is persistent, but several exchange rates integratedshow up non-persistent, or weak persistent. In which, the RMB/EUR and RMB/JPY is have themost co-persistent. The result has important meaning for portfolio decisions of exchange rate.更多还原