查詢結果分析
來源資料
頁籤選單縮合
題 名 | 考慮GARCH效果下的尾部指數與風險值應用=Measuring Tail Thickness under GARCH Effect and an Application to Value-at-Risk |
---|---|
作 者 | 林楚雄; 王韻怡; | 書刊名 | 風險管理學報 |
卷 期 | 8:1 民95.03 |
頁 次 | 頁49-70 |
分類號 | 562.38 |
關鍵詞 | 極值理論; GARCH效果; 尾部指數; VaR-x法; Extreme value theory; GARCH effect; Tail index; VaR-x method; |
語 文 | 中文(Chinese) |
中文摘要 | 極值理論 (Extreme Value Theory, EVT) 是目前應用於研究金融資產尾部行為的重要方法。運用極值理論估計尾部指數時,必需假設資料是來自獨立且相同(Independent, Identically Distributed; i.i.d.) 的機率分配。本文考慮報酬序列具有GARCH 效果下,應用Huisman et al. (2001) 所提出的修正Hill 估計式估計尾部指數並應用於風險值的估計。本文實證發現考慮GARCH效果的尾部指數估計值較不考慮GARCH 效果者為小,亦即金融資產的極值分配事實上並非如同不考慮GARCH 效果下所估計的胖尾。本文將考慮GARCH效果所求得的尾部指數應用到風險值的估計,實證結果顯示較不考慮GARCH效果者以及Risk Metrics方法為準確。由以上的實證結果說明研究者在運用極值理論估計尾部指數與風險值時,若先將具有GARCH效果的資料轉換符合i.i.d.,則可以提高估計的準確性。 |
英文摘要 | Extreme value theory is a powerful and fairly robust framework to study the behavior of a tail distribution. The EVT literature is routinely assumed to be Independent, Identically Distributed (i.i.d.) data. However, it is widely agreed that high-frequency financial asset returns are conditionally heteroskedastic, and hence not i.i.d. In this study, we compare how serious an impact this has under GARCH effect on the application of EVT method to tail index. We find that unconditional tail index without GARCH effect is smaller than that ignores GARCH effect. It means that the tail distribution of financial asset returns without GARCH effect is not fat enough than that ignores GARCH effect. It’s the result of our study that the accuracy of Value-at-Risk is improved when the data are transformed into i.i.d. ones before they are estimated. |
本系統中英文摘要資訊取自各篇刊載內容。