投資人是否能利用財務報表資訊來獲取異常報酬一直是財務學上探討的重 要議題。過去研究中大多顯示在台灣市場以資產報酬率或是股東權益報酬率作為 選股訊號,可以獲取優於大盤的績效。然而有文獻指出,資產報酬率與股東權益 報酬率有設計上的盲點,建議可改採用資本報酬率(ROC)或者其他相關的比率作 為選股訊號。 本研究根據資本報酬相關比率將公司排序,排序後買入較大之20%並同時放 空較小之20%,建立一零投資組合,並進一步探討依照這些訊號篩選所買進以及 賣空的投資組合報酬是否有顯著差異。實證結果發現,大部分的資本報酬相關比 率均能預測未來預期報酬,意即可獲取顯著的異常報酬,但在控制了其他變數之 後,只有Cash ROIC2 (定義為[稅前息前淨利×(1−稅率)+折舊] / (固定資產+淨營運 資金))以及Cash ROIC3(定義為(稅後淨利+折舊) / (固定資產+非現金淨營運資金)) 依然對於未來預期報酬有預測能力。本研究更進一步將表現較佳的Cash ROIC2 以及Cash ROIC3進行迴歸分析,結果顯示這兩個變數不僅在統計上或經濟意涵 上有預測能力,且於資本資產定價模型下,可以建構出承受較低系統性風險的資 產組合。
Whether investors can generate abnormal returns by using the information in financial statements is a central question in finance. The literature has shown that the ROA and ROE can predict future expected returns and thereby generates abnormal returns in Taiwan. However, some studies have indicated the disadvantages of ROA and ROE, and suggest other variables, such as returns on capital (ROC) to predict future expected returns. This study examines whether ROC and its related ratios can be used to predict future expected returns. We buy the bigger 20% and short the smaller 20% based on the sorting of the variable ROC, to constructed a zero investment portfolio. Our empirical results show that ROC and its related ratios can predict future expected returns. However, after controlling for other firm characteristics, only Cash ROIC2 (defined as (EBIT × (1 - tax rate) + depreciation) / (fixed assets + net working capital)) and Cash ROIC3 (defined as (net income + depreciation) / (fixed assets + non-cash net working capital)) possess predictive power on future expected returns. Furthermore, Cash ROIC2 and Cash ROIC3 remain statistically and economically significant after including various control variables in the regression analysis. Finally, the results also show that one can construct a portfolio which bears less systemic risk by exploiting the variables Cash ROIC2 and Cash ROIC3 based on capital asset pricing model.