財務報表在資本市場扮演何種角色是會計學者長久以來關注的焦點。Ohlson的研究將實證研究的重心從解釋股價的行為轉向至會計資訊與企業價值關聯的研究。本研究以美國市場為背景，加入分析師對盈餘預測因素，測試剩餘淨利模型實證上的有效性，並且用以檢視是否能以此模型建立超越大盤的投資組合。實證結果顯示發現分析師的盈餘預測作為其他資訊之代理變數以及分析師之長期成長預測捕捉企業未來成長潛質，有助於提升模型對股價的解釋能力。本研究認為成長機會對股價具有正向之影響，現時股價表現反映投資人對未來預期的結果。

Cheng (2005) 曾經由價值的創造（經濟租）與價值的記錄（保守會計）二構面去討論影響剩餘淨利的因素，他的實證結果顯示，就產業層級而言，產業的超額報酬會隨產業的集中程度、產業的進入障礙以及產業的保守會計因子而增加，就公司層級而言，公司的超額報酬則會隨公司的市佔率、公司規模、公司進入障礙以及公司的保守會計因子而增加。為了瞭解前述的主張是否同樣地適用於台灣，本論文參仿其研究程序，探討經濟租與保守會計因子對於我國剩餘淨利的影響。實證結果顯示，除了公司層級之進入障礙因素外，無論產業層級之超額報酬與公司層級之超額報酬，皆未能隨經濟租與保守會計因子而增加。換言之，本研究認為，後續研究者不應過度單純地使用Cheng (2005) 的發現，相反的，未來研究有必要進一步深入探討適用於我國情況的模型。 / Investigating the impact of value-creation (economic rents) and value-recording (conservative accounting) on residual income, Chen (2005) demonstrates that industry abnormal earnings increases with industry concentration, industry-level barriers to entry, and industry conservative accounting factors. The firm-level abnormal earnings increases with market share, firm size, firm-level barriers to entry, and firm conservative accounting factors. In order to explore whether the previous propositions on the determinants of residual income can be applied in our country, this study examines how economic rents and conservative accounting affect residual earnings in Taiwan. In summary, using Taiwan data, I cannot find evidence to claim that researchers in Taiwan can use Cheng’s findings directly. Therefore, it is worth for future works to explore what factors affect residual income in Taiwan.

本研究以Ohlson (1995) model為發展基礎，並將盈餘定義為（１）剩餘淨利及（２）盈餘水準，分別就此二種不同定義下的盈餘，以年度盈餘時間序列來測試其是否符合線性資訊動態假設，及針對各模式對股權價值估計與預測之結果做比較，探討何種評價方法或模式對於估計真實價值、解釋價格及預測報酬之效果較佳。 研究結果發現，我國上市公司之剩餘淨利及盈餘水準時間序列皆符合線性資訊動態假說。剩餘淨利線性資訊動態模型較能正確預估次一期的盈餘。相較於單獨以帳面價值來估計股價，考慮線性資訊動態模型所建立之各評價模型所預測之估計股價皆未能正確預測權益價值及解釋權益價值之波動。在投資策略方面，因剩餘淨利模型之投資績效最為穩定，在該模型之投資策略下，V/P比率愈低（高）之投資組合獲得之平均股票報酬愈低（高），代表股價愈是被低估的投資組合可賺取更多之投資報酬，這說明剩餘淨利模型預測次期投資報酬之能力最佳，投資人可以其做為建立投資策略的參考。 / Based on Ohlson (1995) model, this study specifies earnings variables as both residual income and earnings levels to test the linear information dynamic (LID) models per se and the ability of competing valuation models to value the contemporaneous stock prices. A comparison of future stock return predicting capability of competing models is also conducted. By using both residual income time series and earnings levels time series for examining the issue on the firms listed on the Taiwan Stock Exchange (TSE), the empirical results support Ohlson’s information dynamics. However, when estimated as a time series, the linear information models using either residual income or earnings levels variables provide value estimates no better than book value does alone. From the investment strategy aspect, the superior predictive ability of the residual income valuation model with respect to future stock returns demonstrates that high (low) V/P ratios gets high (low) investment returns. It implies that the underpriced portfolio makes high investment returns. Accordingly, the residual income valuation is good for estimating returns on the following year and is therefore a valuable investment reference..