題名: 美國與台灣的績優股及替代品之動態聯動性分析
其他題名: Dynamic linkages between the U.S. and Taiwan and substitutions in blue-chip stocks
作者: 王俞凱
郭東華
關鍵字: Ljung-Box 檢定
Jarque-Bera 常態性檢定
廣義自迴歸條件異變異數模型
動態條件相關係數模型
Ljung-Box Test
Jarque-Bera Normality Test
GARCH Model
Dynamic Conditional Correlation Multivariate GARCH model
系所/單位: 統計學系, 商學院
摘要: 摘要 本次研究報告探討台灣加權指數與美國費城半導體指數的連動關係以及在半導體產業中替代品與個別股票報酬率的關係。我們選擇台灣加權指數權重佔比最大的台積電(TSMC)做為代表性個股,而費城半導體指數我們選擇英特爾(Intel)以及超微(AMD)為代表性個股。首先,使用單變量GARCH模型來配適這三檔股票報酬率,預測其波動以及比較三者估計波動的變化。再者,為了要觀察連動性,我們使用了多變量DCC GARCH模型來估計配對股票報酬率的條件相關係數,探討估計結果與替代效應的關係。分析結果發現在相關性方面英特爾與AMD分別對台積電的條件相關係數皆分布於0.2~0.3,呈現正相關,雖然相關性不高但可以知道兩檔股票報酬率相關趨勢;在超微與英特爾的條件相關係數方面,由於這兩間公司為競爭的公司,經分析結果發現彼此的條件相關係數為正的,在2009年時條件相關係數達到0.7。當一間公司推出新產品時,也就是替代作用發生時,會造成條件相關係數在短暫的時間內變成負的,表示替代品效應會反映在股票報酬率上。
Abstract This research mainly discusses the linkage between the Taiwan weighted index (TAIEX) and the Philadelphia semiconductor index (SOX), the relationship of substitution, and their returns. We choose Taiwan Semiconductor Manufacturing (TSMC), the largest weighted stock in TAIEX, Intel Corporation, and Advanced Micro Devices as our representative blue-chip stocks. First, we fit each return series based on the univariate GARCH model, forecast their volatility, and compare the variation of estimated volatility respectively. In order to understand their linkage, we consider a multivariate DCC-GARCH (Dynamic Conditional Correlation) model to estimate the conditional correlation coefficient of bivariate returns. We thus investigate the relationship of the estimated result and substitution effect. In conclusion, in terms of the bivariate returns correlation, the conditional correlation coefficient range of the Intel Corporation and Advanced Micro Devices to Taiwan Semiconductor Manufacturing are 0.2~0.3, respectively. For Intel Corporation and Advanced Micro Devices, we find that their correlation coefficient is positive through the analysis, because the two companies are competitive. Their correlation coefficient is at least 0.7 in 2009, and when a new product is launched from a company, the substitution effect occurs for this new product. It eventually causes the conditional correlation coefficient to become negative in the short run, presenting. That the substitution effect appear in the counterpart stock’s return.
日期: 2020-11-18T08:00:58Z
學年度: 108學年度第二學期
開課老師: 陳婉淑
課程名稱: 時間數列分析
系所: 統計學系, 商學院
分類:商108學年度

文件中的檔案:
檔案 描述 大小格式 
D0585985108243.pdf1.25 MBAdobe PDF檢視/開啟


在 DSpace 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。