由於金融業與電子業為促使台灣經濟成長的主要產業,使得兩類股票成為投資人熱門地投資標的,如何對兩類股價變動趨勢作精確的預測一直以來是投資人與研究者欲探討的議題。而投資人運用台灣股價加權指數作為觀察股價走勢的指標。再者,總體經濟為造成股價波動之重要因素,因此,若我們能充分掌握這三者的資訊,以及納入總體經濟指標對三種股價指數之影響,並建立預測模型,將有助於投資人或管理者進一步制定投資策略。因此本研究以台灣股價加權指數、電子類股股價指數及金融類股股價指數為對象,採用17項總經指標,例如:SEMI半導體接單出貨比、工業生產指數、景氣領先綜合指數、景氣同時綜合指數、外銷訂單數、貨幣總計數M1B等。研究範圍包含三種股價指數之季節趨勢之分析,以及季節指數預測;研究期間從1999年1月到2012年12月,共14年156期之三者每月平均股價指數。使用X-12 ARIMA自我迴歸模型分析三者股價之波動在總體經濟環境下是否有顯著之季節穩定性及移動性,並建構三個變數之季節指數預測模型。經實證分析,得到以下結論: 一、在季節指數穩定性測試之下,三者之股價在總經變數影響下具有淡旺季皆有顯著性差異。 二、在季節指數移動性測試之下,台灣股價加權指數及電子類股之淡旺季在總經變數影響下具有明顯移轉現象;而金融類股淡旺季移轉現象較不明顯。 三、以X12-ARIMA時間數列模型預測三者之季節指數,其MAPE皆小於2.2%,因此準確度理想,具實務應用意涵。
Because electronic Industry and financial Sector play an important role in economic growth in Taiwan, Electronic Industry and financial Sector became the top investment target because they both promote Taiwan’s economy in recent years. Investors and researchers have been always concerned about precise prediction for the price variation of electronic stocks and financial stocks. Besides, the TAIEX is one of useful index for investors to observe the tendency of stock price, macroeconomics is the critical element cause stocks to fluctuate .Therefore, if we can master these three stocks’ information under macro-environment, constructing the forecasting model, it will help investors or managers to establish investment strategies. Hence, we take TAIEX, electronic stocks and financial stocks as our study object, and adopt 17 macroeconomics measurements. Our studies include seasonal trend and the prediction of the seasonal index in the future, and the study period is from January 1999 to December 2012. We use X-12 Autoregressive Model not only to analyze whether the fluctuation of these three components’ stock price will have significant difference in seasonal stability and seasonal moving under macroeconomic factors, but also to construct the forecasting model of three factors’ seasonal indices. Via the empirical test, we summarize three conclusions thereinafter: 1.With the seasonal stability test, there is significant difference in dull and peak season of three components’ stock price under macroeconomics. 2.With the seasonal moving test, there is significant moving phenomenon in dull and peak season of three components’ stock price under macroeconomics. 3.Forecast the future stock price of three factors with X-12 Model. The result show that all three factors’ MAPE are lower than 2.2% , meaning the predictability is high.