中文摘要：

针对上海银行间同业拆放利率（SHIBOR）动态特性中存在的机制转换及波动聚集现象，分别将机制转换和广义自回归条件异方差（generalized autoregressive conditional heteroskedasticity,GARCH）效应引入CIR模型，构建了带机制转换的CIR模型（RSCIR）以及带机制转换和GARCH效应的CIR模型（RSCIR—GARCH），并基于SHIBOR利率数据对CIR模型、RSCIR模型及RSCIR—GARCH模型进行对比分析，发现机制转换和GARCH设定的引入大幅提高了模型的数据拟合优度，并消除了利率波动中的ARCH效应．基于RSCIR—GARCH模型对SHIBOR市场的风险溢价动态特性进行研究，发现SHIBOR市场利率波动显著存在高利率高波动以及低利率低波动两个机制，相应的风险溢价的动态特性也发生了机制转换．

英文摘要：

In the light of regime switching and volatility clustering in the dynamics of SHIBOR, regime- switching CIR model （RSCIR） and regime-switching GARCH CIR model （RSCIR-GARCH） are established by introducing regime-switching and generalized autoregressive conditional heteroskedasticity（GARCH） spec- ifications into CLR model successively. Then, a contrast study among CIR, RSCIR and RSCIR-GARCH models is performed based on SHIBOR sample data, which indicates that the regime-switching and GARCH speci- fications improve the model fitness significantly and eliminate the ARCH effect in the volatility dynamics. Furthermore, an empirical research is carried out on the risk premium dynamics of SHIBORbased on the RSCIR-GARCH model, and it is found that there are two regimes, that is, the regime of higher interest rate with higher volatility and the regime of lower interest rate with lower volatility, in the dynamics of the term structure and risk premium of SHIBOR.