財務數據資訊內涵之比較--以臺灣上市公司為例

The Comparison of Information Content from Financial Data--The Evidence of Taiwan Stock Market

本研究使用 Collins， Maydew and Weiss（1997）的股票評價模式結合每股營業收入、營業利益率與毛利率，以台灣上市公司之 Panel Data 分析財務數據與股價之間的關係，並考慮時間與產業之因素。此外並結合 Biddle、Seow and Siegel （1995）提出資訊內涵的概念，以 Chan and Seow（1996）所提出的 JA 檢定來判斷哪一個資訊最值得讓一般投資人關注。 我們使用模型進行多重迴歸得知，每股淨值的 t 統計量大於每股盈餘，其中以全部樣本、2008 年第三季到 2012 年第三季與傳產狀況類似，t 統計量看起來明顯大很多，符合Collins， Maydew and Weiss（1997）的實證結果，而 2006 年第一季到 2008 年第二季、電子、服務及其它情況類似，t 統計量看起來略大一些。 相對資訊內涵的部分，JA 檢定實證結果發現，在全部樣本每股淨值與每股盈餘的資訊內涵相等，且會比每股營收的資訊內涵多。2006 年第一季到 2008 年第二季與電子的情況相似，每股淨值與每股盈餘的資訊內涵相等，並依序大於營業利益率、毛利率與每股營收。2008 年第三季到 2012 年第三季、傳產、服務及其它的結果一樣，每股淨值、每股盈餘、毛利率與每股營收的資訊內涵相等，並大於營業利益率。 增額資訊內涵的部分，擁有越多資訊並不一定會有越多資訊內涵，我們發現以每股淨值與每股盈餘為基礎，當增加每股營收這項數據時，並沒有增加資訊內涵，反而使資訊內涵減少，所以新增的資訊有可能會與原有資訊互相干擾，導致原有的資訊內涵減少。

In this study， we extend the stock assessment model of Collins， Maydew and Weiss（1997）with revenue per share， operating income percentage and gross margin percentage to investigate the relationship between stock price and financial data for Taiwan listed companies. We also take into account the time and industry factors for our study. On the other hand， we use the JA-testproposed by Chan and Seow （1996） to determine which information deserve the most investor’s attention.Empiricalresults are as follows: From multiple regression models for the whole sample， we learned the book value per share’s T-statistic is greater than the earnings per share’s. There’s similar situation for the third quarter of 2008 to the third quarter of 2012 and traditional industries， which are in line with Collins， Maydew and Weiss （1997） empirical results.But such comparison is not rigorous; it cannot represent the book value per share information content are greater than the information content of earnings per share. For the relativeinformation content JA-test empirical results show that the information content of Book values per share are equal to of earnings per share are equal in the whole sample， and more than the revenue per share. First quarter of 2006 to the second quarter of 2008， and electronic industry have the similar situation， book value per share and earnings per share have the equal information content， and in turn greater than the operating income percentage， gross margin percentage and revenue per share. Third quarter of 2008 to the third quarter of 2012， traditional production， services and other have the similar situation: Book values per share， earnings per share， gross margin percentage and revenue per share have the equal information content. For the incremental information content JA-test empirical results show: The more financial data do not necessarily have more information content. When increasing the revenue per share data， it did not increase the information content to book values per share and earnings per share， but to make the information content offset from each other. Additional financial data may interfere with the original financial information ， and result in the reduction of the information content.