| Summary: | 碩士 === 國立高雄第一科技大學 === 財務管理所 === 95 === By analyzing the trend of short term interest rate in Taiwan, Japan, Hong Kong, Singapore, and South Korea, this paper investigates the issue that whether the extreme value distribution is time dependent at a given threshold. In addition, we examine the existence of GARCH effect in short term interest rate, i.e. the volatility-clustering phenomenon. Together with a moving window approach, the Peak-over Threshold (POT) model based on generalized Pareto distribution is used to estimate the local maximum and minimum extreme values. Empirical evidences indicate that：(1) In all five countries, the first-order differenced series of short term interest rate display leptokurtic distributed and exist a volatility clustering phenomenon. (2) The term structure of interest rate volatility has a downward trend in Taiwan CP market. (3) In testing the existence of time dependent for location parameters from 30, 90 and 180 days Taiwan CP data, we find it has a negative relationship. (4) In testing the existence of autocorrelation for location parameters from Taiwan CP data, the evidences indicate that the 180 days CP follows the AR(1), however, 30 and 90 days CP does not. (5)The extremes of short term interest rate indeed exist the GARCH effect in 90 days Taiwan CP data.
|
|---|
