臺灣地區高科技產業股票報酬率之特性試探

Abstract

在眾多研究台灣股價關聯性的文獻中發現─以基本分析方法所採用的自變數各有不同，亦使用過許多計量經濟模型及各種統計方法，惟針對高科技產業類公司進行分析之文獻較為罕見。因此，本研究試圖以基本分析的觀點來探討影響台灣高科技產業股票報酬率的若干因素；由於影響股價變動的因素不勝枚舉，因此本研究首先從許多研究股價關聯性文獻當中，歸納出經常被引用成為影響股價變化的自變數，並兼顧該變數資料的影響性及取得的便利性。提出以四個總體經濟變數─工業生產量成長率、利率變動率、M2現金供給量變動率、通貨膨脹率及公司四個財務比率─每股盈餘、流動比率、負債比率、存貨週轉率等八個變數，為本論文之研究變數。其次試圖以有別於各文獻所使用的統計方法─panel data(縱橫資料)，亦稱為橫斷面時間序列混合資料，並以固定效果模型及隨機效果模型作為統計分析方法。將所得實證結果相互比較，並與文獻之結果互為印證。 本研究以民國八十五年至八十九年間五十家電子類上市及上櫃公司股票季報酬率與各變數之季資料，經過不同計量模型實證分析結果為：在四個財務比率中，以每股盈餘(EPS)和股票報酬率呈現顯著正相關，其餘如流動比率、負債比率及存貨週轉率則呈現不顯著的關係。而在四個總體經濟指標中，以工業生產量成長率、M2現金供給量變動率呈現顯著正相關，利率變動率呈現顯著負相關，至於通貨膨脹率則呈現不顯著關係。 其次本研究發現：一、以基本分析的角度(以上八個指標變數)研究股價變化相關性，則無論高科技產業或非高科技產業其結論均相同。亦即高科技產業股票報酬率與八個自變數(即工業生產量成長率、利率變動率、M2現金供給量變動率、通貨膨脹率、每股盈餘、流動比率、負債比率、存貨週轉率)的關聯性，無異於非高科技產業股價報酬率與八個自變數的關聯性。二、高科技公司的財務報表所提供之資訊，對高科技公司的股價具有影響力；尤其以〝獲利能力〞一項最為密切，故提高公司的盈餘，即可提高公司的股價。三、總體經濟指標亦影響高科技公司股價變動。因此，倘若擔心外在的經濟景氣產生變化而無法立即顯現在公司財務報表上時，從總體經濟指標去預測高科技公司股價變化仍具意義。

The analytic methods used in literature on discussing stock prices were quite similar. The independent variables in those analyses were all different. Also varied were their econometric models. Besides, it is so seldom to conduct an analyses specifically targeting on high-tech manufacturing companies in Taiwan. These motivate me to develop the basic analytical view on investigating the factors which might affect the stocks of movement for Taiwan's high-tech manufacturing businesses . Since numerous factors were regarded by different sources, we were first to infer the most important independent variables. Surely we had to take into the consideration both their informative influential characters and convenience of accessibility. We end up with four macroeconomic variables, i.e. productivity growth of the whole industry, variation in interest rate, fluctuation in M2 cash supply, and overall inflation rate, and four company-related financial ratios, i.e. equity per share (EPS), liquidity ratio, debt ratio, and inventory turnover rate. The next step we took was to employ a rather new statistical method. The one we chose was something called panel data (vertical and horizontal information), also called cross sectional time series mixed information. As to the statistical analytical method, we picked both a fixed effect model and a random effect one to compare the actual results gathered under each and also with those previously reported for verification. It was mainly employed to assure the consistency of the conclusion if any. This study use of the actual seasonal returns of stock equity of fifty listed and OTC companies in the electronics category and their respective variable information during the period between 1996 and 2000. Through the factual analyses, the results are: among the four company financial ratios, EPS alone shows significant positive relationship with the stock return, but not the other three factors. Whereas among the four macroeconomic indexes, both productivity growth rate and M2 cash supply give positive relations with the stock prices, but interest rate shows an opposite trend and inflation seemed played an indifferent role. We also conclude : First, all stock prices seem to follow a similar trend no matter whether they belong to the high-tech industry or not. Second, information provided by financial reports of high-tech companies has great impacts on their stock prices, especially the item of “profitability.” Third, some macroeconomic indexes also have profound influences on the fluctuation of high-tech stock prices. All in all, we think the stock prices are quite predictable to some degree.