本文利用縱橫平滑轉換迴歸 (panel smooth transition autoregression, PSTAR) 模型評估股價報酬率及其持續性 (persistence) 效果,其中的轉換變數是以景氣對策信號分數衡量之,因為該變數可以作為判斷景氣榮枯與評估國家經濟政策的執行成效的重要參考依據。換言之,本研究考慮落後期的股價報酬率為解釋變數,且以景氣對策信號分數作為轉換變數,估計PSTAR模型。實證上,以46家台灣50指數成分股報酬率為對象,樣本期間為2004年1月至2013年2月,共計5060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,並提供政策決策單位與企業主採取配套措施以穩定股市的依據。 實證結果可歸納如下: 一、 台灣50成份股股票報酬率與其自我落後期項之間存在顯著的非線性關係,且此關係隨轉換變數--景氣對策信號分數不同而變動。換言之,這意味著當期的景氣對策信號分數以非線性型態影響未來的股票報酬率。 二、 當景氣對策信號分數超越其門檻值13.3223時,股票報酬率具有平穩的持續性效果。反之,則股票報酬率為震盪波動的持續性。 三、 由於股票報酬率的持續性介於-46%~37%之間,表示影響股票報酬率的因素,主要仍來自當期總體的基本面與其它訊息,例如:當期利率與匯率等,且其影響比重高於以往歷史報酬率資料。
This paper adopts panel smooth transition autoregression (PSTAR) models to evaluate stock returns and their persistence effects. We consider total scores of monitoring indicators as the transition variable in the PSTAR model, since the indicators are the referenced index for assessing a country’s business cycle and operating performance. That is, we choose lagged stock returns as the regressors and total scores of monitoring indicators as the transition variable to estimate the PSTAR model. In empirical, we select 46 component companies of the Taiwan 50 Index as the sample objects. The sample period spans from January 2004 to February 2013. Thus, there are 5060 observations. Our empirical results provide useful information for the investors in the stock markets, the managers of enterprises, and the authority of financial markets. The empirical results can be summarized as follows: 1. The stock returns of the component companies of the Taiwan 50 index display a nonlinear dynamic path, depending on the transition variable: total scores of monitoring indicators in different regimes. That is, current total scores of monitoring indicators nonlinearly influence future stock returns. 2. When the total scores of monitoring indicators are over the threshold value, 13.3223, the persistence effect is smooth; however, as the total scores of monitoring indicators are below the threshold value, the persistence effect is fluctuant. 3. Since the persistence of stock returns ranges from -0.46 to 0.37, implying that there are still some crucial macroeconomic factors influencing the persistence of stock returns. For example, current interest rate, foreign exchange rate, and price level.