| Summary: | 碩士 === 東海大學 === 財務金融學系 === 99 === This study attempts to explore the dynamics between investor sentiment and portfolio returns of Taiwan stock markets under different frequencies data (day, week, month, quarter, year) with the control of Fama and French three-factor model (market, market size and book to market value). We implement the empirical results by adopting the multivariate ADCC-GARCH model to capture the information transmission between investor sentiment and portfolio returns and volatility dynamics under different data frequencies. Empirical results show that investor sentiment help to predict portfolio returns of different frequencies, providing the insights for investors under decision making process. Among them, the very short-term (days) operating investors, when investors fear sentiment rises, the excess returns can obtained by the momentum strategy; the other hand, for a longer operating period (week, month quarter) investors, the more investors fear, the contrarian strategy operations provide more opportunity to make abnormal returns. Furthermore, under the operation of momentum and contrarian strategies, we found the unexpected return provides important influence on the volatility of investor sentiment. In addition, the investor sentiment shows the symmetric effect on the volatility of portfolio returns, indicating the increased level of investor fear have more significant impact on the volatility of portfolio returns than that of lower level of investor sentiment. Finally, the portfolio returns and investor fear demonstrate the significant persistent information on covariance movement; moreover, the previous negative portfolio returns and higher investor fears significantly increase the co movement on next period.
|
|---|
