- 學位論文
投資情緒對股價報酬持續性的影響—縱橫平滑轉換自我迴歸模型之應用
The Effect of Investment Sentiment on Stock Returns Persistence:An Application of Panel Smooth Transition Auto-Regresssion Model
摘要
摘 要 本研究利用縱橫平滑轉換自我迴歸(panel smooth transition autoregression, PSTAR)模型探討股價報酬率及其持續性(persistence)效果。在PSTAR模型中，採用股價報酬率的自我落後期當作解釋變數，恐慌指數(Volatility Index，VIX)當作轉換變數，估計恐慌指數對股價報酬產生的門檻效果，以及評估股價報酬的持續性。實證上以43家台灣半導體上市公司的股價報酬率為對象，樣本期間為2003年1月至2013年12月，共計5676個觀察值。實證結果可以提供股市投資者進行股票投資時的參考，並提供政府決策單位採取配套措施以穩定股市的依據。 實證結果可歸納如下: 一、 43家台灣半導體上市公司的股票報酬率與其自我落後期之間存在顯著的非線性關係，決定於各期恐慌指數的水準值，且其結果決定於轉換變數恐慌指數(VIX)具有2個不同的門檻值，分別為23.3227與26.2878。 二、 當投資人的情緒(VIX)處於低度恐慌及高度恐慌時，股價指數報酬率之持續性皆低於中度恐慌，顯示當VIX介於23.3227~26.2878時，股價指數報酬率受到當期外在干擾的影響小於VIX位於其他區間。 三、 股市在面對重大的經濟與金融事件時，恐慌指數會上升，對股價造成暫時性衝擊，而使股價報酬的持續性反轉為負值。
並列摘要
Abstract
This thesis employs a panel smooth transition autoregression (PSTAR) model to evaluate stock returns and their persistence effects. In the PSTAR model, we use lagged stock returns as the regressors and the volatility index(VIX) as the transition variable to find the threshold effect of the VIX on stock returns and the persistence of stock returns.
In empirical, we select 43 semiconductor companies listed on the Taiwan Security Exchange Corporation as the sample objects. The sample period spans 2003 1M to 2013 12M.Thus,there are 5676 observations. Our empirical results provide useful information for the investors to make investment strategies in the stock market and for the authority of financial markets to make or revise relative policies for stabilizing stock market.
The empirical results can be summarized as follows:
First, the relationship between stock return and lagged stock returns for the 43 sample companies in non-linear,depending on the VIX in three different regimes. The threshold values of the VIX are 23.32227 and 26.2878.
Second, the stock returns persistence is the biggest as the VIX locates in the middle regime (i.e.,23.3227