透過您的圖書館登入
IP:18.189.193.172
  • 學位論文

投資組合風險值估算模型之探討-多變量MAR-GARCH模型

Estimation of Portfolio VaR – Multivariate MAR-GARCH Model

指導教授 : 吳祥華教授
共同指導教授 : 李孟峰副教授(Mong-Hong Lee)
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


本研究的主要目的在於開發並檢測新的風險值測量工具,使得一般投資人與機構投資人能有效率的評估投資組合所暴露的風險,進而管控投資組合的風險,在能承擔的風險水準下尋求最大的獲利。本研究利用台灣證券交易市場中金融股價指數與電子股價指數當作樣本資產,原始資料取自「TEJ台灣經濟新報資料庫」,資料型態為日資料,研究期間為1995年1月7日至2007年6月23日,共3266筆資料。本研究延伸了Wong and Li(2001)、Lanne and Saikkonen(2003)與柯婷玲(2004)的單變量MAR-GARCH 模型,在考慮資產間共變異的特性下,建議以多變量MAR-GARCH之波動估算模型,並進一步利用該模型估計應用至預測投資組合的風險值。 根據實證結果發現,傳統雙變量GARCH模型僅在99%信賴水準下通過檢定,樣本外的檢定效果並非十分理想,而本研究所發展出的雙變量-MAR-GARCH由於考量到結構轉變的特性,發現Lanne and Saikkonen雙變量-MAR-GARCH模型樣本外在95%信賴水準與99%信賴水準下皆通過檢定,且Wing and Li雙變量-MAR-GARCH模型樣本外則全部通過檢定,由此可知雙變量-MAR-GARCH模型有效的修正了傳統雙變量GARCH模型沒有考慮到結構轉換的問題,雙變量-MAR-GARCH模型確實有著較佳的預估風險能力。 因此,根據實證發現,多變量MAR-GARCH模型 不但能考慮到投資組合可能處於不同狀態,對不同狀態下的模型進行估計,多變量的設定更能有效的捕捉投資組合中資產間的共變異特性,相較於傳統雙變量GARCH模型,以及單變量MAR-GARCH模型,多變量MAR-GARCH模型確實能有著較佳的預估風險能力。

並列摘要


This thesis mainly develops and tests new VaR measurement model, multi-variables MAR-GARCH model, for investors to estimate the risk of portfolio efficiently. With this new model, investors can maximize the profits under the control of portfolio risk. Wong and Li(2001) and Lanne and Saikkonen(2003) introduce single-variable MAR-GARCH model. In nowadays, VaR measurement only for single-asset is inappropriate. Under considering the covariance structure of portfolio assets, we urge multi-variables MAR-GARCH model to estimate the portfolio VaR. The data we used are collected form Taiwan TEJ database, which include Financial index and Electronic index in Taiwan stock market from 1995/1/7 to 2007/6/23. We use equal weights of these two assets to form a studying portfolio. With in-sample and out-sample test, the empirical results show that multi-variables MAR-GARCH model not only take state-variable into consideration, but also estimate parameters within different states. The multi-variable modeling can capture the covariance between portfolio assets in different sate. Comparing to traditional bi-variate GARCH model and single-variable MAR-GARCH model, multi-variables MAR-GARCH model indeed have better predict ability to estimate the VaR of portfolio in our empirical evidence.

參考文獻


[2] Baillie, R. T. and R. J. Myers,(1991),“Bivariate GARCH Estimation of the Optimal Commodity Futures Hedge”, Journal of Applied Econometrics, Vol. 6, pp.109-124.
[3] Bautista, C. C.,(2005)“How Volatile are East Asian Stocks During High Volatility Periods?”, Applied Economics Letters, 12, pp.319-326.
[4] Bollerslev, T. (1986),“Generalized Autoregressive Conditional Heteroskedasiticity”, Journal of Econometrics, 31, pp.307-327.
[5] Bollerslev, T. (1987), “A Conditionally Heteroskedasticity Time Series Model for Speculative Prices and Rates of Return”, Review of Economics and Statistics, 19, pp.542-547.
[6] Brooks, C., and Persand, G., (2003), “Volatility Forecasting for Risk Management”, Journal of Forecasting, Vol.22, No.1, pp.1-22.

被引用紀錄


楊朝仲(2011)。風險係數應用於我國壽險業之研究-次級房貸風暴前後之比較研究-〔碩士論文,淡江大學〕。華藝線上圖書館。https://doi.org/10.6846/TKU.2011.00765
朱慧培(2012)。以時變自我迴歸模型預測金融商品之風險值〔碩士論文,國立臺北大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0023-2807201213024800
林素杏(2016)。應用DEA交叉效率評估在投資組合的選取-以我國股市為例〔碩士論文,國立臺中科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0061-1907201608493000

延伸閱讀