Summary: | 碩士 === 國立臺北大學 === 合作經濟學系 === 92 === The views by past scholar for futures and foreign exchanges which affected stock markets have some controversy. There is no consensus because those researches haven’t simultaneously incorporated the information about futures and exchange rate, however, to compare across the markets collectively is more important. That is, only giving emphasis to analyze single market must result in fallacious statistics and serious error inference. In order to make up a deficiency and correct the mistakes, this research employed multivariate error correction GJR GARCH-M with constant representation model to describe the characteristics of financial data, such as skewness, fat-tailed, leptokurtosis and non-normal distribution. The model process provide the stock investors in Taiwan to verify the dynamic linkage in short-term, mean and volatility spillover among stocks, futures and foreign exchange market in Taiwan, U.S. and Japan, including investigation on risk premium, volatility asymmetry or volatility clustering, the linkage of across markets and lead-lag situation, and long-run equilibrium. Eventually, the model can be used to simulate the changes in the financial markets and specifically provides the effective investment decision-making suggestion.
The empirical results present that as the integration of regional economy increased, the market correlation among Taiwan, United States and Japan is positive significant and through the linkage of foreign exchanges, the stock rewards of them are connected more closer. Secondly, by the recognition of lead-lag relationship among these markets, the US stock market influents the stock markets of Taiwan and Japan via the transmissions of returns and volatility in the future and stock markets in many ways and the US stock market is only affected by both of the reward rate and volatility return itself. Both on these two aspects, the US no doubt demonstrates its leading position in the world financial market. Similarly, in Asia market, the evidence of Japan stock market influents on the future and stock markets of Taiwan is obvious. On the other hand, the Taiwan stock market influenced by US and Japan markets is relatively weak. Besides, the conditional average of these countries only effects by the returns of their previous term and without risk premium. Meanwhile, the long term equilibrium of Taiwan future and stock markets will be achieved quickly in next period after the short term biases occurred. Moreover, in conditional variance, the own volatility of these markets is the main cause of these variations and the unexpected negative and asymmetric volatility spillovers of US stock market also cause severer variation on other markets. However, in stock market, there isn’t uniformly volatility which means the market response by negative impacts is generally greater than that of positive impacts. Finally, the diagnose test for the fitness of multivariate error correction GJR GARCH-M with constant representation model could eliminate the serial correlation for error in first-moment and second-moment which also capture the ARCH and GARCH effects and well fitness in volatility clustering, the asymmetric volatility in its or across markets and returns spillover. This research also concludes the rewards and fluctuations of exchange markets in the conditional average and variation functions to demonstrate the improvement of the explanation of the model. The analytic results of the thesis also provide advantageous information for the government authorities in investment and financial policies making.
|