Measuring extremal distribution of spot rate risks with extreme value theory and copulas
| Autor: | Yi-Neng Liao, 廖怡能 |
|---|---|
| Rok vydání: | 2006 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 94 The data this paper use are Taiwan interbank call loan rate from July, 1, 1982 to June, 30, 2005 because they can be sensitive to the recreasing of US federal funds interest rates. Spot rate will bring large damage when it gets large volatility. This paper analyze the fluctuations of maximal and minimal changes distribution in spot rate. This research calculates VaR when the extreme event happens and compares different kinds of models. This research combines GARCH model and extreme value theory for estimating the VaR and ES because spot rate has heteroskedasticity and autocorrelation. This paper uses the concept of AR-GARCH-copula to measure VaR and ES of portfolio which is composed of Taiwan interbank call loan rate and US federal funds interest rates. This paper fit extreme value copulas to extreme value distribution of spot rate portfolio and compare the effect of the different copulas. The empirical findings indicate spot rate is right skewness, high kurtosis distribution and the existence of volatility clustering, fat tail distribution, heteroskedasticity and autocorrelation. Finally, this paper find AR-GARCH-GPD model to give particularly accurate Value-at-Risk measures to spot rate extreme distribution and get a good simulation to spot rate portfolio extremal distribution. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|