摘要
以33个行业股票组合为分析样本,应用CU SUM SQ统计量对β系数进行稳定性检验,结果发现中国行业股票组合普遍存在β系数不稳定性特征。采用状态空间模型直接对β系数的时变行为进行建模,平均绝对预测误差M AE和平均平方预测误差M SE表明,基于卡尔曼滤波的市场模型有更好的预测效果,中国行业股票组合系统风险系数β的时变性行为可用均值回复过程来描述。
This paper aims to investigate the form of systematic risk of Chhinese industrial stock returns. Thirty three industrial betas were tested by using CUSUMSQ statistic and fund that thirty of them are unstable over time. Time-varying betas were estimated using recursive regressions, rolling regressions and two classes of state-space models based on the Kalman Filter. comparing the performance of these models with the mean absolute forecasting error and the mean square forecasting error. The result suggests that the Kalman filter approach is preferred and the variation of industrial portfolio beta can be described using mean-reverting process.
出处
《系统工程》
CSCD
北大核心
2006年第2期62-67,共6页
Systems Engineering
