本研究旨在觀察投資人情緒對台灣股價指數報酬率的影響,所選用的情緒代理變數來自於現貨、期貨及選擇權市場的交易指標,包括了ARMS、券資比、法人的台指期貨多空交易量比、以台股加權指數為標的的賣買權交易量比和未平倉量比、以及選擇權反推的股價波動率指數。加入了台幣對美元的匯率變數後,本文將情緒指標做不同的排列組合,以向量自我迴歸(Vector Autoregressive, VAR)模型估計變數間的關係。除了台灣股票指數報酬率外,本文亦分析了情緒變數對寶來台灣卓越50證券投資信託基金(簡稱台灣50)報酬率、和不同產業的股價指數報酬率的影響。 結果顯示,投資人情緒對各種股價指數的影響不完全相同,其中選擇權反推的股價波動率指數及法人的台指期貨多空交易量比對台股指數、台灣50及各產業股價指數的報酬率皆有顯著的影響,又加入選擇權隱含的股價波動率指數的模型其配適度均比未加入者高,顯示關於「價格」的情緒變數對報酬率有較好解釋能力。然而,因本文使用的是日資料,能考量的總體變數有限,儘管投資人的情緒對股指報酬率常有顯著的影響,但模型的整體配適度仍偏低。
This thesis tries to investigate the impact of investor sentiment on returns of Taiwan stock indices. The proxy variables for investor sentiment are from the spot, futures and option markets. The sentiment proxy variables include: ARMS, ratios of short sells to margin purchases (SP), trading volume ratios of long to short TAIEX futures by financial corporations (TIF), trading volume ratios of put to call (PC), open interest ratios of put to call (OPC) and volatilities of indices returns implied by option premia (TVIX). After adding exchange rates between NTD and USD, vector autoregressive (VAR) models are employed to estimate relationships between stock indices returns and various combinations of investor sentiment variables. In addition to TAIEX returns, this study also considers Polaris Taiwan Top 50 Tracker Fund returns and individual industry indices returns. The results show that impacts of investor sentiment variables on distinct indices returns are dissimilar. Variables of TVIX and TIF are found to have consistent and significant influences on all stock indices returns. The fitness of VAR models including variable TVIX is always better than that for models excluding it. This suggests that price-related investor sentiment variables may be more influential on stock indices returns. However, although investor sentiment variables have significant impacts on stock indices returns, the overall fitness of VAR regressions in this study is pretty low because daily data are used and limited macroeconomic variables can be considered in the models.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。