透過您的圖書館登入
IP:3.147.85.183
  • 學位論文

匯率風險值衡量之實證研究-以新台幣、日圓、英鎊、歐元匯率為例

Study on the Measurement of Foreign Exchange Risk: VaR Measurement and Back-testing on New Taiwan Dollar,Japanese Yen,British Pound and Euro

指導教授 : 張傳章
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


自一九七O年代初期國際貨幣體系進入了浮動匯率時代,近三十多年來,匯率對一個國家之經濟活動具有重要的影響,匯率過度波動不僅為經濟帶來不利影響,還會影響商品的價格而導致價格扭曲。傳統之匯率風險管理方法是著重在總體經濟的研究,但是匯率的變動並非僅以傳統的總體經濟的基本理論來解釋,於是近年來新的匯率風險管理方式,已成學術界、金融業甚或企業的重要課題。 本文將風險值概念及衡量風險值之模型作一個簡單介紹後,主要著眼於外匯市場風險值問題的探討,分別依不同之風險值評估方法計算風險值(Value at Risk ,VaR),並用不同之回溯測試方法檢定不同之風險值評估方法與實際發生之風險其中之誤差,藉由其異質性來探討各個不同之風險值評估方法對匯率波動之適用性。 本研究結論如下: 1.整體而言,本研究中除美元兌換台幣之匯率外,指數加權移動平均法(EWMA)及等權移動平均法(MA)多優於歷史模擬法與蒙地卡羅模擬法。 2.指數加權移動平均法(EWMA)在中長期風險檢測有不錯之表現,主要是假設過去的資料與鄰近的資料對參數估計的效果不同,λ值愈小,代表最近的觀察值愈能包含最多的資訊。 3.歷史模擬法之誤差除了美元兌換台幣外則最大,有極端情況產生時極易低估風險或高估風險,這跟歷史資料之取樣有關,所以在本文第三章第二節中提及補救之方法:即指數加權移動平均法(EWMA)及拔靴複製法(Bootstrapping Methods)以增加其檢測風險之準確性。 4.蒙地卡羅模擬法,蒙地卡羅模擬法基本上是基於大數法則的實證方法,當其模擬之次數越多,愈能將各種較極端之風險情形考慮進去,其平均值會越趨近於理論值,在本研究中其結果在日圓、英鎊及歐元與MA較接近,另外在台幣方面則與EWMA較接近。

關鍵字

風險值 外匯市場 匯率風險

並列摘要


This thesis is focusing on the study of measurement of VaR in the foreign exchange(FX)market. We use four different VaR evaluation models to measure the VaR, and then use three different back-testing methods to verify which the relatively best model will apply for in different FX markets with respective confidence level and window length. We conclude this study as listed below: First, Generally, the EWMA method and MA method are relatively better methods than historical simulation method and monte carlo simulation method, except for the USD against TWD fx market. Second, For the EWMA method, it usually performs well for the VaR measurement in the mid-term(150 days)and long-term(250 days)FX markets, due to the presuming that the nearer and farther data in history use different parameter λ, the smaller λ represents coverage of more latest data. We use 0.94 as λ in this study. Third, for the historical simulation method, the FX risk will be devaluated or over-valuated in the extreme situations. This is because the outcome of the study depends on what the sampling of the historical data we take. In the section 2 of the chapter 3 in this thesis, we have mentioned two methods to make up or minimize the differences in measuing VaR, those are EWMA method and Bootstrapping method. Fourth, For the monte carlo simulation method, the more times we simulate, the more extreme situations will be considered. In this study, we find that the results are very close to MA in JPY, GBP and EUR FX markets. On the other hand, it is very close to EWMA in the USD against TWD fx market .

參考文獻


[7] 陳佩鈴(2002)「匯率條件風險值之估計與比較」,中原大學國際貿易研究所,碩士論文。
[2] 何中達編譯(2005),國際財務管理(Multinational Business Finance)- David K. Eiteman, Arthur I. Stonehill and Michael H. Moffet原著,初版,
[1] Alexander, C. and C. Leigh, (1997), “On the Covariance Matrices Used in VaR Models”, Journal of Derivatives, pp.50-62,
[2] Beder, T. S. (1995), “VaR:Seductive but Dangerous” ,Financial Analysts Journal, Vol. 51, (September-October), pp. 12-24.
[3] Bollerslev, T., (1986), “Generalized Autoregressive Conditional Heteroskedasticity”, Journal of Econometrics, 31, pp.307-327.

延伸閱讀