Based On The Dynamic Behavior Of The Regime Switching Model Of Short-term Rates

The dynamic behavior of short-term interest rates plays a very important role on fixed-income securities and interest rate derivatives. Since understand of the dynamics of the short-term interest rates can help to discover the prices of financial products, manage the risk of interest rates. Short-term interest rate is an important business cycle variable, which is important for the formulation of monetary policy and the sensitivity of inflation expectations. Through analysis of the dynamic behavior of short-term interest rates, macroeconomic policies can be effective to adapt to economic cycles and changes in inflation expectations, evaluation of the effects of economic policy. Therefore, a large number of domestic and foreign scholars conduct research on short-term interest rates. But most of the studies based on Brownian motion to describe the dynamic behavior of short-term interest rate, so the sample path of short-term interest rate is always continuous. However a large number of empirical studies found that the sample rate is not always coutinuous, there will be structural changes in interest rate behavior. In fact, short-term rates in China actually undergo significant structural changes and in terest rate volatility has obvious boundaries. Therefore, it is necessary to take into account the structural changes in short-term interest rates to capture the time-varying state of the economic changes.This study focus on three aspects. Firstly, to extend the single factor interest model-CKLS based on the regime switching model. The regime variable follows a continuous time two-state Markov chain.Secondly, intruducting the details of a theory established by Ait-Sahalia (2002) to approximate the true density function of an arbitrary univariate diffusion process in closed form. Then highly accurate transition density function of thediffusion process is used when constructing the likelihood function by applying the recursive algorithm developed by Hamilton (1989). Therefore, our study does not suffer from discretization bias. Finally, conducting our empirical analysis of short-term interest rates through maximum likelihood estimation and comparing our model with models of Vasicek and CIR.The maximum likelihood estimates are obtained usingthe daily series of China seven-day inter-bank lending close rate. The estimation results reveal that there are strong evidences for the existence of high and low volatility regimes, for the time varyingtransition probability of the regime variable, and for the high persistence of both regimes. In regimes, the volatility, but not the drift, is estimated accurately and plays a key role in explaining the dynamics of the interest rates. High persistence's and different volatilities of two regimes can well explain volatility clustering observed in the data. Based on the inferred probability of the process being in each regime, most of the high volatility periods correspond to some historic events. The likelihood-based test shows that misspecification can result in misleading outcomes particularly regarding the volatility and transition probabilities of the regime index.