基于可变强度跳跃-GARCH模型的资产价格跳跃行为分析——以中国上市公司股票市场数据为例

近年来，美国金融危机、欧债危机、地震等突发事件不断冲击着我国金融市场，各类资产价格频繁出现大幅跳动，收益风险短期内急剧扩大。鉴于此，本文构建了门限效应下状态变量依赖自回归强度跳跃-GARCH模型（简称TSD-ARJI-GARCH模型）来探讨股票资产价格随时间平滑波动和大幅度跳跃的双重特征。该模型扩展了现有可变强度跳跃-GARCH模型，克服了片面强调内生或外生因素的局限性，既允许跳跃强度受单个资产异质因素的内生驱动，以刻画跳跃变化的时变性及集聚性，也考虑了外部状态变量影响的门限效应。通过对不同类型中国上市公司股票市场数据的实证分析，验证了该模型对各类上市公司股票资产价格跳跃特征都具有较好的辨别和预测能力，可为动态监管金融资产的跳跃风险提供理论依据。

Analyzing the Jump Dynamics of Asset Price in Jump-GARCH Model with Variable Intensity

Recently, the unexpected events, such as the US financial crisis, the EU debt crisis and the earthquakes, constantly shock China's financial markets. Under these impacts, almost all sorts of financial assets suffer frequent jumps in their prices, which then cause much higher risk in the short period of time. In order to describe both of the normal volatility and the jump change in financial asset, a mixed jump-GARCH model,that is TSD-ARJI-GARCH model,is constructed in this paper. It extends the existing models to make the jump intensity not only driven by internal idiosyncratic factors, but impacted by external state variables. This model takes into account time variation, clustering and threshold-based state-dependence of jump change as well as asymmetry and clustering of normal volatility. Then, stock price data of listed companies in China's stock market is used to test these different models. Both goodness of fit and forecast evaluation show TSD-ARJI-GARCH model performs better. Therefore, it can provide relatively better theoretical and technological support for monitoring and managing the jump-risk of financial assets in practice.