透過您的圖書館登入
IP:3.15.221.67
  • 學位論文

金融控股公司市場風險值之模型---對稱與不對稱模型之比較

Modeling Value at Risk of Financial Companies--- A Comparison of Symmetric and Asymmetric Models

指導教授 : 蘇永成
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


近幾年來,台灣的金融控股公司快速的成立,並且廣泛地、多角化地進行擴張,在這個環境之下,隨著頻繁且大量的交易活動的產生,金融控股公司所面臨的市場風險也逐漸增大,為了控制金控公司所曝露的風險,以及遵守巴塞爾協定的規定,我們必須妥善做好市場風險的衡量。 本論文採用GARCH家族中的兩種模型…ARMA(1,1)-GARCHM(1,1)與ARMA(1,1)-EGARCHM(1,1),這兩種模型分別假設殘值〈random shock〉對稱與不對稱,我們利用這兩個不同假設的模型,對自行模擬的兩個投資組合進行風險值〈VaR〉的預測,並對兩種模型的預測結果比較其有效性及效率性。 此論文研究發現,過去的殘值〈random shock〉對於條件變異數存在著不對稱的影響,並且我們所採用的不對稱模型…ARMA(1,1)-EGARCHM(1,1)具有較佳的預測能力,提供金融控股公司在市場風險管理上的有效性與效率性。

並列摘要


In recent years, the financial holding companies in Taiwan have established rapidly and expanded widely and diversely. On the circumstance, the gradually significant market risk accompany with the frequency and large volume of the trading activities. In order to conform to the Basle 1996 Amendment and to measure the trading risk exposure, we adopt two models of GARCH family, ARMA(1,1)-GARCHM(1,1) and ARMA(1,1)-EGARCHM(1,1) models, to forecast the VaRs of the two simulated portfolios, and compare them on the basis of the effectiveness and efficiency. This paper provided the concept on the asymmetric effect of the past random shocks on the volatility and proposed the well-forecasted ability in EGARCHM(1,1). This finding provides the best VaR-forecasting model which is effective and efficient to financial holding companies in Taiwan.

參考文獻


5.Basak Suleyman and Alexander Sharpiro, 2001, “Value-at-
risks. Bank for International Settlements, 1996b.
4.Barone-Adesi Giovanni, Giannopoulos Kostas and Les
Vosper, 1999, “VaR without Correlations for Portfolios
of Derivative Securities”, The Journal of Futures

延伸閱讀