摘要
本文基于A股市场的高频数据,利用阈值估计量对资产的连续波动和跳跃进行区分,分别估计出包含跳跃(RCOV)和不包含跳跃(CCOV)的协方差矩阵,并利用HAR-CJ模型对两种协方差矩阵进行样本外预测,最终将预测的协方差矩阵应用于全局方差最小策略和风险平价策略。研究发现:当持仓周期较短时,考虑跳跃可以获得更好的投资表现,反之当持仓周期大于等于两周或更长时间,不考虑跳跃的投资组合表现更好;在市场剧烈波动时,跳跃影响投资组合的持续性更长,而在市场平稳时,跳跃影响投资组合的持续性会减弱。
Based on the high-frequency data of the A-share market,this study estimate and forecast the covariance matrices with jump(RCOV) and without jump(CCOV) respectively,and finally applies the predicted covariance matrix to reveal the effect of discontinuous volatility(Jump) on the performance of the portfolio strategy.The empirical results find that:1.When the holding period is short,better investment returns can be obtained by considering jump variation;conversely,when the holding period is greater than two weeks or longer,the portfolio that does not consider jumps performs better.2.When the market fluctuates violently,the impact of jump on the investment portfolio is more persistent;and when the market is stable,the impact of jump on the investment portfolio is less persistent.
作者
陈梓荣
马征程
郑旭
Chen Zirong;Ma Zhengcheng;Zheng Xu
出处
《投资研究》
北大核心
2023年第6期22-37,共16页
Review of Investment Studies
基金
国家自然科学基金资助项目“基于高频数据的Levy跳跃扩散模型的计量检验和估计”(项目编号:72073096)。
关键词
价格跳跃
投资组合
持仓周期
市场环境
Price jump
Portfolio management
Holding period
Market environment