Vega加權平均與ATM選擇權隱含波動率於 CreditGrades 模型之比較

Abstract

本研究參考 Finger (2002)、Stamicar, & Finger (2006) 所發表之 CreditGradesTM 模型,因認為價內外選擇權在流動性良好之情況下,皆包含有價 帄選擇權所未包含之市場資訊。故將原文獻利用價帄 (At-The-Money; ATM) 選 擇權隱含波動率改為使用 Vega 加權帄均隱含波動率,再與原文獻方法所得之預 測值作誤差分析比較。 本研究使用 CreditGradesTM 風險評估模型,並選定 2009 年到 2013 年間,美 國 9 間公司股票及選擇權流動性良好之上市公司,用以衡量樣本公司之信用風險, 並比較利用不同隱含波動率導入模型之結果,檢驗何種隱含波動率較適合用於 CreditGradesTM 風險評估模型,並且探討是否在 2008 年金融海嘯過後,此模型 對於信用違約交換 (Credit Default Swap ;CDS ) 之報價預測能力是否準確。 經本文透過不同隱含波動率求得方法比較與個案分析之實證結果發現,利用 CreditGradesTM 模型估計所估計出之理論信用利差與實際信用利差走勢一致,且 利用 Vega 加權帄均隱含波動率所得到之結果優於原模型利用價帄選擇權隱含波 動率之結果。

This research uses CreditGrades Model which was published by Finger (2002) and Stamicar, & Finger (2006). Both the In-the-Money options and the Out-the-Money options provide some market information that are not included in the At-the Money options. Thus, this research replaced the ATM implied volatility by Vega average weighted implied volatility. This research uses CreditGradesTM Risk Evaluation Model and chooses data of 9 public companies with great liquidity in their stocks and options in U.S. from 2009 to 2013. I input two implied volatilities with different definitions to CreditGradesTM Risk Evaluation Model to test their capability of generating accurate result when being applied to CreditGradesTM Risk Evaluation Model. Moreover, we will discuss the accuracy of the forecasting ability of the CreditGradesTM model by comparing the estimated values to the actual CDS spreads generated by data after financial crisis in 2008. Through the empirical research, I compare these two different implied volatility methods and case studies for each sample company. I find that not only the trends of the estimated value and actual value are consistent, the research outcome of adopting the vega weighted average implied volatility is better than the ATM implied volatility.