應用GARCH-EVT-Copula模型於外匯投資組合風險值之評估

Applied GARCH-EVT-Copula Model to Estimate the VaR of Exchange Portfolio

賴政宏(Cheng-Hung Lai)

指導教授 ： 李沃牆

淡江大學/商管學院/財務金融學系碩士班/碩士(2015年)

https://doi.org/10.6846/TKU.2015.00825

摘要

本研究以APEC組織成員國中與台灣貿易頻繁的國家─中國、香港、新加坡、美國、日本、韓國等六國所組成的外匯投資組合為例，且資料區間考量中國第一次匯改政策時點，故取2005年7月21日起至2014年7月21日，共九年之資料；而研究方法本文有別於以往傳統定義報酬率尾部分配為常態的假設，而以GARCH、極端值理論、Copula函數所建立的GARCH-EVT-Copula非線性模型來衡量該投資組合的風險值。實證方面，建立等權重的投資組合，並且比較傳統VAR-COV模型與靜態、動態的GARCH-EVT-Copula模型所估算投資組合的風險值，再以回溯測試及LR檢定尋找最適模型；本文資料區間橫跨金融海嘯的極端事件，然實證結果顯示，靜態GARCH-EVT- Frank Copula模型所估算的風險值可有效應用於本文外匯投資組合之避險，無論是否發生極端事件，都能有效達到該投資組合避險的效果，可供台灣出口商作為避險之依據。

關鍵字

GARCH ； 極端值理論 ； Copula ； 風險值

並列摘要

The work focuses on risk management of foreign exchange portfolio and it consists of the nations who often trade internationally with Taiwan. Additionally, those nations are also the member of Asia-Pacific Economic Cooperation, including China, Hong Kung, Singapore, USA, Japan and South Korea. However, the work also consider the date that the government of China changed its exchange rate policy on July 21st in 2005. That’s the reason we decided to accept the data of exchange rate from that day to July 21st in 2014.
We applied nonlinear static and dynamic GARCH-EVT-Copula model to measure downside risk rather than assumes returns are normally distributed. To compare with the empirical results in our work, we also measure value at risk (VaR) with the variance-covariance method. Subsequently, we use back-testing method and likelihood ratio test to figure out which model could evaluate VaR accurately.
In consequence, static GARCH-EVT-Frank Copula model is the optimal model for the exchange portfolio in the work. Even extreme events exist during the nine-year period, our optimal model still out performed than the others. We also suggest that exporters or investors in Taiwan could manage the risk of exchange portfolio with our model.

並列關鍵字

GARCH model ； Extreme-value Theory ； Copula function ； Value at Risk

參考文獻

曾智業 (2012)，國際投資組合之風險值評估，淡江大學財務金融研究所碩士論文。

黃泰源 (2013)，應用Copula函數於金磚五國投資組合相關性及風險值評估，淡江大學財務金融研究所碩士論文。

Allen, D. E., A. K. Singh, and R. J. Powell, (2013), “EVT and Tail-risk Modelling: Evidence from Market Indices and Volatility Series,” North American Journal of Economics and Finance, Vol. 26, pp.355-369

Artzner, P., F. Delbaen, J. Eber, and D. Heath, (1999), “Coherent Measure of Risk,” Mathematical Finance, Vol. 9, pp.203–228

Beder, T. S., (1995), “VAR: Seductive But Dangerous,” Financial Analysts Journal, Vol. 3, pp.12-24