波動度作為一風險衡量工具，在強調財務風險管理的今日倍顯其重要。本研究選擇七種不同波動度模型進行比較，包括：40日歷史波動度、VXO指數、價平等加權平均隱含波動度、等加權平均隱含波動度、Vega加權平均隱含波動度、成交量加權平均隱含波動度及VIX指數，期找出符合台灣選擇權市場現況之波動度指標。研究發現： 一、各波動度模型與未來真實波動度間的誤差衡量，會隨著所納入的資訊內涵愈 多，其誤差愈小；而且，各波動度模型間彼此的差距也同樣會隨天數的增加 而逐漸縮小。 二、VXO對未來市場真實波動度較其他模型更具預測能力。 三、VXO與大盤同期報酬呈負向相關。 四、VXO與同期報酬呈現不對稱關係，關係方向則隨觀察天數不同而不同。 五、大盤短期報酬對VXO有顯著之規模效果，該效果隨報酬天數增加而減弱。 六、VXO對大盤未來報酬率較其他模型更具預測能力。 七、鑑於低VXO水準之相對未來報酬顯著為負；高VXO之未來報酬顯著為正； 隱含不同VXO水準時的買進賣出策略。穩定度分析顯示，前述VXO對未來 真實波動度較其他模型更具預測能力之結論，相當穩健。 上述發現有助於投資者發掘最適波動度指標，形成最適投資決策。分析過程亦發現傳統評估波動度方法有其限制，進而提出建議。 As an important risk indicator, volatility becomes more prominent since ever due to the emphasis on risk control. This study nominated volatility estimation models including 40 days-HV, VXO, ATM, EWIV, VWIV, TVIV and VIX for the comparison aimed on optimal volatility index in TXO. The analyses list as following : 1. The error measurement between all volatility models and real volatility will become lower by containing more information. 2. VXO preferably predicts the real volatility than others. 3. VXO showed a negative-contemporaneous relation with the market return. 4. VXO showed an asymmetric relation with the return. 5. VXO showed a scale effect with the short run return, such an effect decays with the term increases. 6. VXO preferably predicts the return than others. 7. Whereas the low VXO matched with significant negative future return, the high VXO matched with significant positive future return; hints investors the adequate sell out or buy in strategies when facing with different benchmarks of VXO. The robustness test tells aforesaid finding of that VXO preferably predicts the real volatility than others quite firm. The findings help investors to excavate both the optimal volatility index and investment strategy in TXO. Furthermore, this study emerges the deficiency of traditional criteria in valuing the volatility model thus some advices were made.
