摘要:選擇權市場的主要功能之一是提供重要的經濟變數及事件的資訊,以作為經濟政策及投資的參考依據,有很多文獻都提到選擇權市場具有股票市場動向的資訊,例如Easley, O’Hara, and Srinivas (1998)的研究發現買入賣出的選擇權交易量可以預測未來的股票價格,顯示出具有資訊優勢的投資人會在個股選擇權市場交易。隨著個股選擇權市場的快速成長,個股選擇權市場的實證研究變得越來越重要,最近Chang, Christoffersen, Jacobs, and Vainberg (2011)提出了一個方法可以利用個股選擇權的價格資料來計算個股的CAPM β值,由於β是財務上非常重要的一個變數,所以本研究打算檢驗選擇權市場隱含的β值是否能夠即時的反映系統風險的變化。目前既有的文獻都是利用市場模型以及過去的歷史資料來估計β值,這些傳統方法計算出來的β值並沒有辦法有效解釋橫斷面的股票報酬資料,Lewellen and Nagel (2006)更進一步指出即使允許β隨時間變動的,條件CAPM模型仍然無法有效捕捉資產定價模型中的異常現象。相反地,我們認為選擇權隱含的β值具有市場價格風險溢酬的資訊,因此較傳統的CAPM模型的β值能更顯著的被定價,本計畫將檢定這個假說。此外,文獻也指出個股選擇權能改進市場效率性,亦即選擇權市場的資訊會擴散到現貨市場,並提前反映股票市場的資訊(例如購併的消息或者是盈餘宣告等)。為了檢驗個股選擇權的價格資料是否具有訊息內涵,我們打算研究Hou and Moskowitz (2005)的價格延遲溢酬,以及意外的盈餘宣告值是否能夠被選擇權的隱含β值所解釋。本研究計畫也將重新檢驗Goyal and Saretto (2009)的橫斷面的個股選擇權報酬研究,該研究將個股的歷史波動度及個股選擇權的隱含波動度的差距排序,買進差距最大的前10%的個股選擇權,然後放空差距最小的10%的個股選擇權,結果發現這樣的操作策略會產生非常高的超額報酬,此超額報酬沒有辦法被傳統的因子模型所解釋(例如Fama and French (1993)的三因子模型),本研究打算重新探討這個議題,我們打算採用Kraus and Litzenberger (1976)的系統共偏態模型,我們認為這個模型可以解釋橫斷面的個股選擇權報酬。最後,我們打算研究CBOE利用指數選擇權與個股選擇權的價格所計算的隱含相關係數,CBOE指出當股票市場下跌時隱含相關係數會上升,因此隱含相關係數可能具有解釋與預測重要經濟變數的功能,因此本研究打算研究議題。此外,我們也將探討Driessen, Maenhout, and Vilkov (2009)提出的相關係數交易策略的績效,探討它與隱含相關係數的關係等。
Abstract: With the rapid growth of individual stock option market, the empirical analyses of this market become an important research topic. Recently, Chang, Christoffersen, Jacobs, and Vainberg (2011) have proposed a method to compute CAPM betas using the individual stock option prices. In this sub-project, we will first will examine whether option-implied CAPM betas reflect the systematic risk in a timely fashion. As pointed out by Jegadeesh and Titman (1993) and Fama and French (1992), the cross-sectional stock returns cannot be explained well by historical betas. Furthermore, Lewellen and Nagel (2006) also show that even with time-varying beta, conditional CAPM cannot capture asset-pricing anomalies such as the momentum and the value premium. In contrast, we expect that option-implied beta can be priced much more significant than historical CAPM beta.
Moreover, Easely, O’Hara, and Srinivas (1998), Chakravarty, Gluen, and Mayhew (2004), Cao, Chen, and Griffin (2005), and Pan and Poteshman (2006), provide evidence that information flows from option markets to stock markets, and informed investors exploit their private information in the option market before in its underlying asset market. Thus, the information is disseminated from the option market to the underlying asset market. Thus, we expect that option-implied beta may have the information advantage especially during major corporate events, such like earnings announcements, and merger and acquisition (M&A) announcements, in which case options market disseminates information proceeds stock market and historical betas. To test the above argument, we will exam whether the price delay premium, documented by Hou and Moskowitz (2005), can be explained by option-implied beta, and whether the option-implied beta can reflect firm-specific information such as earnings surprise.
This project will also reinvestigate the finding of Goyal and Saretto (2009) that a zero-cost trading strategy, which long (short) the portfolio with a large positive (negative) difference between historical realized volatility and at-the-money implied volatility, produces an economically and statistically significant average monthly return. Their results are robust to different market conditions, to stock risks-characteristics, to various industry groupings, to option liquidity characteristics, and are not explained by usual risk factor models. Goyal and Saretto (2009) suggest that the profits of their long-short portfolios may arise as compensation for some unknown aggregate risk. Under this case, formulating a cross-sectional options return model that accounts for their portfolios’ returns is an important future research. Inspired by the work of Vanden (2006), this project will test whether market coskewness risk augmented factor model of Kraus and Litzenberger (1976) can explain the cross-section stock option returns. Moreover, this project will also examine whether the coskewness risk is priced in cross-section of option returns.
Finally, this project will investigate the information contents of CBOE’s option implied correlation. For example, CBOE has documented that, similar to VIX, implied correlation exhibits a tendency to increase when the S&P 500 decreases. Besides the CBOE’s option implied correlation, Driessen, Maenhout, and Vilkov (2009, p.1390-1391) have constructed a correlation trading strategy. It is worth linking the performance of the correlation trading strategy to the option implied correlation. We can further test whether the correlation risk is a risk factor priced by stocks and/or options. Moreover, It is widely documented that the investors have the “flight to quality” behavior during the financial crisis period. Therefore it is worth studying whether the option implied correlation can predict the occurrence of the “flight to quality” behavior.