衍生性金融商品的資訊內涵整合型研究-子計畫四---選擇權隱含資訊，交易策略與流動性提供

Abstract

本計畫為一個三年期計畫，在探討選擇權所隱含之價格與風險訊息之資訊內涵，主要 著重於隱含波動之分析，將探討其預測績效與對交易策略與流動性提供之影響。本計 畫先從何謂已實現波動(或真實波動,realized volatility, RV)加以探討，分析近期所討論關 於市場微結構與價格跳躍因素對已實現波動之影響。本計畫將運用 Andersen, Bollerslev, and Dobrev (2005), Barndorff-Nielsen and Shephard (2005), Christensen and Nielsen (2005) 所提出檢測跳躍因素之方法，解析日內高頻率價格資料中跳躍因素之效 果，進而對已實現波動拆解成連續性的累積波動成份與跳躍因素之波動成份。將先嘗 試比較多種檢驗Jump 成份的檢定方法之效力。在預測波動度之模型上，本研究除選 取傳統之Black Scholes implied volatility (BSIV)，將運用由Britten-Jones and Neuberger (2000)所先提出，後由Jiang and Tian (JT, 2005) 更進一步分析的model free implied volatility (MF-IV)，JT 證明MF-IV 在diffusion process 有Jump dynamic 時仍然是適用 的，本計劃將據此討論MF-IV, BS-IV, lag RV 等之預測效果。並進一步解析三種模式對 連續性的累積波動成份與跳躍因素之波動成份的預測效果。本計畫先以OptionMetrics 資料庫中選擇權指數與個股選擇權資料為樣本，除進行隱含波動與歷史已實現波動之 一般預測效果檢定之外。有別於過去研究多只針對單一指數或外匯或個股選擇權加以 分析，本研究將採取較大之個股選擇權以橫斷面模式藉由Panel Data model 來解析，跨 股票間各種模式預測效果差異性之原因。本研究之第二、三年，將此類問題作深入跨 入另一相關領域之分析，其中包括交易策略分析與選擇權市場流動性提供之分析。由 於國內衍生性商品近年來快速發展，本計畫也將進一步以國內選擇權契約為基礎加以 分析，檢驗BS 隱含波動之預測能力是一個驗證選擇權與股市效率性與BS 模式的定式 是否有問題之聯合檢定，由於我國股票市場中不完美因素較多，例如漲跌幅與放空限 制等，MFIV 可能顯現出較佳之預測效果，但MFIV 可能又有數值積分(numerical integration)估算誤差等問題；因此本研究將藉由實證分析來探討各模式在台灣市場之可 運用性。本計畫之分析主題與子計畫(1), (3), (6)與(7)等等子計畫都有極大相關，希望發 揮團體合作之綜效，完成高品質之學術論文投稿國外財金類一流期刊。

The study of market volatility has been one of fastest growing areas in financial research, as understanding volatility in capital markets is important for evaluating investment and asset allocation decision, for pricing derivative assets and for determining the cost of capital. In this research we will first explore the information content of various implied volatility model. We will compare the performance of Black Scholes implied volatility (BS-IV) to an alternative model free implied volatility (MF-IV) originally proposed by Britten-Jones and Neuberger (2000). Jiang and Tian (2005) provide a more detail study on it. The model free IV requires more inputs than the model based IV and the more the prices of all options having the same maturity and the more strikes available the better. Our research investigates the information content of BS-IV and MF-IV for the forecast of realized volatility. Using the technique provided by Barndorff-Nielsen and Shephard (2005), we decompose the realized volatility into jump variation and continuous integrated volatility components, which are further studied in our analysis. The empirical study will be based on the TAIFEX options, CME options on S&P 500 futures, and OptionMmetrics database. In the second year, this project will examine the economic benefits of using high frequency volatility measure, implied volatility, and the traditional time series volatility model such as GARCH and GJR-GARCH. We will also discuss the possible method of optimally combining information from two dimensions of volatility information: implied and realized volatility. The risk and return of trading straddle and other volatility trading strategies is further explored. The analysis will cover the economic value of volatility trading based on implied volatility, time series model of realized volatility and Jump-starting GARCH models. Alternatively, the performance of various volatility models are compared by their economic benefits of derivative pricing and hedging for the case that financial institutions writing options and conducting dynamic delta hedging. In the third year this research will consider the effects of implied information on option liquidity. Existing studies have established extensive evidence on the time-varying bid/ask spreads in the market microstructure research. This research takes a different approach and explores a previously unexamined feature of implied volatility by investigating the time-series properties of option bid/ask spreads. The main purpose is to analyze the relationship between call and put spreads and the information shock, i.e., the volatility. We investigate whether BS-IV and MF-IV provide incremental information contents on the liquidity of option markets.