利用平滑化隱含波動度曲線預測資產報酬

Abstract

過去不少文獻的結果顯示隱含波動度曲線的偏離程度以及買賣權的波動度差值對於股價走勢有良好的解釋能力。本研究試圖改善過去僅使用特定執行價與現貨價比值的資料來建構因子，我們先為波動度資料建立平滑的曲線，再依此建立修正的指標，並與過去文獻的偏離指標一同比較。研究結果顯示負偏向的波動度曲線與標的資產的報酬有正面的相關性，而此特性隨著不同的時間區間而有變化。此外，曲線的模型參數也有額外的解釋力，顯示此模型具有一定的貢獻力。

Many prior studies have shown that the skewness and spread of implied volatility curve contains the information of predicting future asset returns. This thesis aims to improve the construction of the option-based measures which traditionally utilize only implied volatilities pertaining to specific moneyness levels. A parsimonious model is proposed to fit implied volatility curves, and next we construct the tradition measures with the parsimonious model. Our empirical results indicate that higher asset returns are positively correlated to the more negatively skewed implied volatility curves and those relationships are robust in different examined in periods. Besides, the best-fitted parameter values of our parsimonious model can provide the extra predictive power, which implies our model captures the useful forward-looking features from option markets.