本研究觀察我國盈餘與股價的關係，共測試八個模式其中包括價格模式、報酬模式與ERC模式。本研究以為(1)ERC模式優於其他模式，(2)ERC模式下盈餘是有用的(3)盈餘持續性、成長機會與風險為盈餘股價關係模式中重要的變數，惟加入實證模式中，結果會較不穩定。(4)未預期盈餘與持續性應以成長率形式定義，及(5)報酬窗期應交明確定義之。茲將實證結果彙述如下：
1.本研究的實證結果顯示ERC模式下的盈餘與股價關係確實優於其他的價格模式與報酬模式。
2.實證結果顯示我國盈餘確具資訊內涵，且越往後的年度，其資訊有用的程度越高。
3.實證結果亦顯示在ERC模式中，因加入盈餘持續性、風險及成長機會，雖然模式解釋增加，惟結果的變異程度加大。
4.以往的研究通常多使用報酬模式，但從經濟的直覺上來說，價格模式所估出的斜率係數應該是較不偏倚的，基於這種直覺，本研究以實證來支持價格模式較不偏倚的理論。
當股價並未在盈餘之前先行反映時，則無論是價格模式或報酬模式都能有不偏的係數，（預期為1/r），惟本研究是假定股價在盈餘之前先行反映(prices lead earnings)，此時，價格模式將有一不偏的盈餘反應係數，另一方面，報酬模式則將產生一較偏的盈餘反應係數。本研究結果額示，價格模式確能有一較不偏的係數估計值。但係數的標準誤則較大，統計顯著性上較差，兩另一方面，儘管報酬模式的標準誤較小，統計顯著性較佳，但卻是一個含有較為偏倚的係數估計值。 / Return models and price models were commonly employed by previous research. This paper makes a comparison between these models, on considering the specific circumstances of Taiwan stock market in which majority, always up to 90%, of the stockholders are short-term individual speculators. In view of long-run earnings formulation different from single period one. It also specifies earnings response coefficient factors including persistence, growth opportunity, and risk, as the independent variables.
The empirical results confirm that:
1. After considering the long-run effects of earnings response coefficient factors, the models including the ERC factors do dominate simple regression, despite the increase in variance.
2. This study also confirms that Earnings do contain information contents, and the significance is increasing over the years.
3. Price models are better specified in that the estimated slope coefficients from price models, but not return models, are unbiased, and this results is consistent with economic intuition. However, return models have less serious econometric problems than price models.
| Identifer | oai:union.ndltd.org:CHENGCHI/B2002002482 |
| Creators | 楊陳松 |
| Publisher | 國立政治大學 |
| Source Sets | National Chengchi University Libraries |
| Language | 中文 |
| Detected Language | English |
| Type | text |
| Rights | Copyright © nccu library on behalf of the copyright holders |
Page generated in 0.0021 seconds