• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

以厚尾分配及緩長記憶特性模型分析日圓匯率期貨報酬之風險值 / VaR Analysis for the Dollar/Yen Exchange Rate Futures Returns with Fat-Tails and Long Memory

鄭士緯, Cheng, Shih-Wei Unknown Date (has links)
本篇文章將採用長期記憶模型之一的HYGARCH模型,搭配1985年廣場協議後的日圓匯率期貨資料來估計日圓期貨匯率買入和放空部位的日報酬風險值,探討控管日圓匯率期貨在使用上的風險。為了更準確地計算風險值,本文採用常態分配、學生t分配以及偏態學生t分配來作模型估計以及風險值之計算。 本文實證的結果將有兩方面的貢獻:首先,實證結果顯示當我們採用厚尾分配估計風險值時,樣本內風險值的估計誤差會與信賴水準的高低呈正比的現象,證明在極端的風險值估計上,厚尾分配均有較佳的表現。其次,與其他使用HYGARCH模型研究日圓匯率的文章相較,本文在風險控管層面上所提供的偏態學生t分配,於估計風險值時,比起只考慮厚尾的對稱學生t分配將來得更為有效,其不但在估計誤差上較小,而且根據Kupiec檢定法,其在樣本內的風險值估計也有較好的表現。此外,本文也將多方證明此資料的偏態分配屬於右偏。 / In order to manage the exposure of the dollar/yen futures returns with regarding the long memory behavior in volatility, we use the HYGARCH(1,d,1) model with the data after the Plaza Accord to compute daily Value-at-Risk (VaR) of long and short trading positions. To take into account the fat-tail situation in financial time series, we estimate the model under the normal, Student-t, and skewed Student-t distributions. The contribution of this article is twofold. First, the empirical results show that the bias of in-sample VaR increases as the confidence level increases when VaR is calculated with a fat-tail distribution. Second, we provide a better distribution, the skewed Student-t innovation, for estimating the HYGARCH model for the Japanese yen in respect of risk management because the bias under the skewed Student-t innovation is smaller than that under the Student-t distribution, and in-sample VaR of the models with a skewed Student-t distribution outperforms based on Kupiec test. In addition, we get the innovation skewed to the right through the in-sample VaR analysis.

Page generated in 0.0214 seconds