Abstract: In view of nonlinearity,policy impacts,big volatility and possible structural changes in Chinese short rate,this paper extends the Markov-regime switching and stochastic volatility model proposed by Smith (2002) and introduces nonlinearity to drift and markov switches to all coefficients in stochastic volatility equation. Using Chinese 7-day interbank offered rate data,we find there exists obvious nonlinearity,regime switching and level volatility effects in the rate,and obvious volatility persistence reduction after taking the regime switching into account. Additionally,high-regime probabilities imply high volatility and high inflation rate,low-regime probabilities imply low volatility and low inflation rate.

Key words: short rate, Markov-regime switching, stochastic volatility, Kim filter