连续波动的累积变化是否触发随机跳跃?——来自国际股票市场的证据  

Do the cumulative changes of volatility trigger jumps? Evidence from international equity markets

作　　者：朱福敏[1] 刘仪榕 郑尊信[1] ZHU Fu-min;LIU Yi-rong;ZHENG Zun-xin(College of Economics,Shenzhen University,Shenzhen 518060,China;Business School,University of Nottingham Ningbo,Ningbo 315100,China)

机构地区：[1]深圳大学经济学院,深圳518060 [2]宁波诺丁汉大学商学院,宁波315100

出　　处：《管理科学学报》2023年第7期54-75,共22页Journal of Management Sciences in China

基　　金：国家自然科学基金资助项目(72071132,72173089)。

摘　　要：股票市场存在跳跃自激发现象和波动率集聚特征已经成为共识,但市场内部随机跳跃和连续波动间是否会相互转化、波动累积是否触发随机跳跃等问题尚存争议.为此,本文将连续波动的累积变化视作价格的“量变”,间断跳跃视作“质变”,采用动态跳扩散双因子交叉回馈模型,借助条件特征函数,引入广义矩估计与粒子滤波方法(GMM-PF),针对具有代表性的国际股票市场进行实证研究,并捕捉和量化两者之间的动态关系.研究显示,随机跳跃和连续波动呈现协同演化、交互传导的现象,一方面随机跳跃将改变下一期的波动率,另一方面量变引起质变,波动率累积也会提高未来的跳跃达到率.同时,与极大似然粒子滤波估计(MLE-PF)、序贯贝叶斯学习方法(SBL)相比,本文提出的GMM-PF方法可实现估计精度与估计效率的联合最优.研究还发现,跳扩散之间的传导机制在不同市场中存在较大差异,针对各个市场的量化结果来看,量变引起质变的程度普遍大于质变引起量变的水平.相比境外发达市场,大多数新兴市场对跳跃风险的消化、转移和分散能力相对较弱,其跳跃集聚和持续性处于较高水平.由此可见,监管当局和投资者有必要对跳跃演变规律给予足够重视.The existence of jump self-exciting and volatility clustering in the stock markets has become a common sense.However,there are still controversies about whether random jumps and continuous fluctuations in the market will be converted into each other and whether accumulated volatility will trigger random jumps.Therefore,this paper treats the cumulative changes of continuous volatility as the “quantitative changes” of prices,and the intermittent jumps as the “qualitative changes”.Then,a two-factor cross feedback dynamic model of jump self-exciting and volatility clustering,combined with the conditional characteristic function of the Lévy process,the generalized method of moments,and particle filtering approaches(GMM-PF),is adopted to capture the interaction between jumps and volatility of important international equity market indexes.Empirical research on representative international stock markets is conducted to capture and quantify the interactive transmission mechanism and degree of jumps and diffusion.Results show that there are co-evolution and cross-feedback effects between random jumps and continuous volatility:Jumps will lead to changes in the continuous volatility of the next period,and the accumulation of volatility will also increase the future jump arrival rate.Hence,the quantitative changes of volatility will cause the qualitative changes of jumps.Compared with the maximum likelihood estimation and particle filtering approach(MLE-PF) and sequential Bayesian learning method(SBL),the GMM-PF method proposed in this paper is more efficient and faster with high precision.Further,this paper also documents that the transmission mechanism between jump and diffusion differs greatly in different markets.Specifically,compared with developed markets,most emerging markets are relatively weak in digesting,transferring and diversifying jump risk,making their persistence of jumps remain at a high level.Therefore,both regulatory authorities and investors need to pay more attention to the evolution of jumps.

关 键 词：量变引起质变 跳跃自激发 跳扩散交叉回馈 广义矩估计 粒子滤波 

分 类 号：F830[经济管理—金融学]