本文探討選擇權的隱含波動度(Implied Volatility)與現貨市場波動度之間是否存在領先落後關係,運用了Hansen and Seo (2002) 兩位學者所提出的門檻向量誤差修正模型(TVECM),對隱含波動度與現貨波動度進行非線性探討,此方法比以往VECM多了門檻值的檢定,將極端值與正常值分為兩個區間,改善單一區間的估計偏誤。本研究中,隱含波動度部分以價性(Delta and K/S)與價格區間(價平(ATM)、價內(ITM)、價外(OTM))搭配方式與現貨市場波動做探討,而現貨市場波動度方面則是利用台灣股價指數分別算出真實波動度(Realized Volatility)、歷史波動度(Historical Volatility)、高低波動度(Daily high/low range)、指數報酬取絕對值(absolute daily Returns)與Garch波動度等五種波動度。TVECM檢定結果發現,多數配對組合存在門檻效果;再由兩區域之Granger causality檢測,發現在大多數組合都呈現相互領先狀態,尤其是在第二區間最為顯著,而第一區間方面結果雖也是互有領先落後關係但則是較趨向選擇權隱含波動度領先現貨市場波動度。此外我們也發現隱含波動度(IV)幾乎都是處於領先現貨市場波動或是相互領先的情況,並無發現有現貨市場波動度單一領先隱含波動度的情況。
This study investigates lead-lag relations between implied volatility for Taiwan index options and stock market volatility. Our study uses two moneyness measures for Delta and K/S to compute option implied volatility under three categories (ATM、OTM、ITM) . Five alternatives of the stock market volatility are realized volatility, historical volatility, daily high/low range, absolute daily returns, and Garch(1,1). Our findings show that there are bidirectional causalities between two variables in most cases. The magnitude of the causality from implied volatility to stock market volatility is larger than the magnitude of reverse relation. It fails to find the evidence that stock market volatility leads implied volatility. Our findings shed lights on lead-lag relations between implied volatility and volatility.