| With the rapid development of international economy, international financialmarkets are becoming closely related and coordinated and tend to be integrated. withthe high-speed development of computer, telecommunication and information technology brings about more and more obvious connection, coordination and noteworthyintegration among international financial markets. During this integration process,individual financial market and that of different types promote and influence each other.Additionally, this integration enhances co-movement among these markets, whichmakes it easy for international financial market risks to transmit among differentcountries and markets, meaning that one volatility in one market leads to volatility inanother market or more markets. With the aim to explore their mutual influencemechanism, risk conduction direction and conduction path, this dissertation choosescrude oil future market and crude oil spot market, American stock market and goldmarket to represent international financial market, adopts the super bayesian influencenetworks theory and the relevant theories and methods in econometrics to deeply studythe dynamic conduction and volatility spillover effects of international financialmarkets.Firstly, this dissertation raises the theory of the super bayesian influence networksand builds a financial information combination framework between dynamicconduction and volatility spillover. Under this framework, mainly introduce thebayesian network, dynamic bayesian network, super bayesian influence networks,co-integration test, vector auto-regression theory and multivariate GARCH model.Secondly, this dissertation studies international financial market dynamic correlation. International financial market correlation is the basis for its risk dynamicconduction and volatility spillover effect. Sampled by crude oil future and spot markets,American stock market and gold market, an econometrics model is built to empiricallystudy dynamic correlation among these market and to discuss their dynamic influenceand co-movement. Subjected by WTI crude oil future market and WTI crude oil spotmarket, S&P500index and gold market, reveal the dynamic correlation among stockmarket, gold market, crude oil future and spot markets, to reveal temporalcharacteristics of the correlation among these markets and to analyze the reasons. Empirical results show that there are significant temporal characteristics of correlationamong WTI crude oil future and spot markets, S&P500index and gold market, andthat the volatility is featured by relatively strong stability, co-movement andconduction among WTI crude oil future and spot markets, stock market and goldmarket. according to the international financial market dynamic correlation empiricalresults, this paper established the graphics of the correlation of the internationalfinancial market network.Thirdly, this dissertation studies dynamic conduction relationship in internationalfin ancial markets. Econometrics methods of co-integration test, vector auto-regression(VAR) model, Granger causal relationship test, impulse response function and varianceanalysis are employed to study dynamic conduction relationship in internationalfinancial markets. It quantitatively describes the dynamic conduction relationship andthe duration of the conductive effect among stock market, gold market, crude oil futureand spot markets. Empirical results reveal that there is long-term stable equilibriumrelationship among S&P500index, gold market, crude oil future and spot markets,that there is relatively obvious dynamic conduction effect among these markets, but ofshort duration, and that the shock received by each market is mainly from itself, but inexternal markets, the largest shock between stock market and gold market is from oneanother, and the same is between crude oil future and spot markets. according to thecontribution of variance decomposition, this paper builds the graphics of the dynamictransmission of international financial market with the super bayesian influencenetworks graphics.Fourthly, this dissertation studies the volatility spillover effect in internationalfinan cial markets. Their mutual influence relationship and volatility spillover effectare studied to seek directions of volatility spillover and information conduction. So,this dissertation builds bivariate GARC H-BEKK model to empirically test volatilityspillover effect among stock market, gold market, WTI crude oil future and spotmarkets, to explore volatility spillover direction and information conduction path ininternational financial markets, and to reveal the origins of volatility spillover effect.according to the direction of the volatility spillover, this paper constructs theinternational financial market directed acyclic super bayesian influence networksgraphics. |
