The Construction And Analysis Of Term Structure Models Of Interest Rates With Regime Switches

The short-term riskless rate is one of the most fundamental and important prices determined in financial markets. It is driving the changes in the entire term structure. This makes the choice of a model for short-term rates crucial to pricing bonds, and pricing interest rate derivatives. The default-free short-term interest rate is also a key economic variable since it is an important input for business cycle analysis through its impact on the cost of credit, its sensitivity to the stance of monetary policy and to inflationary expectations. So it has become one of the most frequently modeled variables in financial economics. Unfortunately, many popular single-regime models do a poor job of modeling the short-term interest rates. Thus this paper develops regime-switching models, in which short-term interest rate is subject to discrete regime shifts that follow a first-order Markov process, and the parameters are different due to different regimes.In more details, firstly, this paper introduces two most common classes of interest rate models, that is diffusion models and GARCH models. Secondly, on the basis of the analysis of regime switches' econometric and realistic meanings the paper analyzes the interest rate in China, and presumes the existence of regime switches. Then under the aforementioned framework of term structure models of interest rates, the paper introduces regime switches and analyzes the characteristics of various regime-switching models. But these models are not so easy to be estimated. To solve the problem, the recursive algorithm is used in simplifying construction of the likelihood function, which results in quasi-maximum likelihood estimates. At the same time, the paper also introduces four kinds of information criteria together with likelihood ratio test in the evaluation of the models. Finally, the empirical research is done in Chinese market. As expected, the results show that regime switching is an important component for improving the model. The restriction on the data may reduce the reliability of the conclusions, but it can become the basis of our ongoing research with the rapid development of Chinese financial market and the gradual adequacy of the interest rate data.