本研究應用選擇權定價模式，以臺灣股票為標的之認購權證為實證分析對象，探 討 Cox （ 1975 ）所提出的 CEV （ Constant Elasticity Variance ）定價模式對臺股認 購權證價格的解釋能力，期望對發行機構及投資人有參考的價值。 本研究將實證分析區分為兩個階段， 在第一個階段中，以成對 t 檢定檢測由選擇權定價模 式所推算出的理論價格，與臺股認購權證的市場價格之間，是否存在著顯著的差異。第二個 階段則以迴歸分析蒐尋造成理論價格與市價有所偏誤的原因。分析構面包括認購權證處於價 內的程度、認購權證距到期日所剩的期間、標的股票的股價波動程度等。研究結果發現： 1. 對於長期居於價內的臺股認購權證而言，選擇權定價模式會高估其理論價格。對於長期 居於價外的臺股認購權證而言，選擇權定價模式會低估其理論價格。但對於持續在價平附近 波動的臺股認購權而言，則不適用上述的結論。 2. 臺股認購權證市價與理論價格的差異， 並不因時間的經過而減少，亦即臺股認購權證可 能不具有市場學習的效果。臺股認購權證市價與理論價格的差異，並不因標的股票報酬波動 的大小而有不同。

　　　　　This paper attempts to employ the CEV pricing model proposed by Cox (1975) to empirically examine the pricing of Taiwan stock market related call warrants. The results will have some referential value for issuing institutions and investors. This study adopts two stages to implement empirical analysis. In the first stage, it performs paired test to investigate the deviation between theoretical prices and market prices. In the second stage, it performs regression analysis to identify the factors affectimg the proce deviations. The factors taken into consideration include the degree of in-the-money、 maturity date、 and stock volatility. The findings can be summarized as follows: 1. Theoretical model tends to overprice those warrants that experience in the money for a lasting period. Theoretical model tends to underprice those warrants that experience out of the money for a lasting period. Conclusions are not suitable to apply to those warrants that are continually fluctuating around at the money for a lasting period. 2. There are no significant evidences to support that pricing deviations between actual prices and theoretical prices will decrease over time. There are also no significant evidences to infer that models overprice warrants when volatility is high. Whether positive or negative, prediction errors of models are on average larger for in the money than for out of the money when warrants experience in the money for a lasting period. In the contradiction, for warrants that are continually fluctuating around at the money for a lasting period, out of the money will cause larger prediction errors.