基于扩展的跳跃SV模型的全国社保基金波动研究

The Analysis of the National Social Security Fund Volatility based on the Extended SV Jump Model

对传统的跳跃SV模型进行扩展,提出了波动率方程中带有协变量的跳跃SV模型,给出了模型参数估计的MCMC算法,并将扩展的跳跃SV模型用于研究全国社保基金的波动特征。研究发现,相对于SV-N、SV-T、SV-M、杠杆SV-N和传统的跳跃SV-N等模型,扩展的跳跃SV模型建模效果最好;社保基金收益率具有明显的跳跃特征,并且跳跃的概率较高;基金重仓的对数ARCH项每增加1个百分点,社保重仓的对数波动率将增加0.5188个百分点。

The traditional SV（stochastic volatility） jump model is extend by adding covariates into the log-volatility equation,and a Markov Chain Monte Carlo（MCMC） algorithm is constructed to estimate parameters.Then the extended SV jump model is used to analyze the volatility character of the National Social Security Fund.It is found that,firstly,the extended SV jump model is much better than the models of SV-N,SV-T,SV-M,leverage SV and traditional SV jump;Secondly,the yield of NSSF has a significant jump character and the probability of jumping is high;Thirdly,when the log volatility ratio of fund Shigekura rises 1%,the log volatility ratio of the National Social Security Fund will rise 0.5188%.