中文摘要：

针对我国股市波动的现实和投资行为的异质性，利用已实现波动率设定包含多个机制平滑转换结构的自回归模型，研究我国股市波动的非对称性及非对称性的非线性动态调整。模型的检验结果发现，我国股市波动存在一个显著的非线性动态调整机制，机制转换的时点发生在收益率的冲击为1．177个百分点处，转换时点的非负性揭示了市场普遍存在的“售盈持亏”行为；股市波动在两个平稳的自回归机制及其之间非线性平滑调整或转换，但转换的概率偏低，速度较慢，自回归过程的平稳性表明投资者具有“过度自信”的决策行为；波动的非对称性没有表现为确定的杠杆效应，在滞后两期的波动水平较高且信息冲击的程度超过阈值时，股市波动表现为反向杠杆效应。以上估计与检验结果基本准确揭示了市场投资者的行为特征。

英文摘要：

This paper studies the asymmetry and its nonlinear adjustment of the volatility in China＇s stock market, based on the reality of volatility and the heterogeneity of investment behavior. Using realized volatility, we propose a multiple-regime smooth transition autoregressive model. The results show that there exists a significant nonlinear dynamic transition regime in China＇s stock market, the transition point of which occurs at the return shocks of 1.177 percentages, and its non-negativity reveals the preva- lence of ＇selling winning stocks and holding losing stocks＇ behavior in the market. Stock market conducts the nonlinear smooth adjustment or transition in two stationary autoregressive regimes. However, the probability of transition is relatively low and the speed is mild. The stationarity of autoregressive process shows that investors have ＇overconfident＇ decision-making behavior. The asymmetric volatility does not display a determined leverage effect, but stock market volatility displays a reverse one when the volatility with two-period lag is high and the information shocks exceed the threshold value. Our results mostly illustrate the properties of investors＇ behavior in the market.