基于时变ARMA-EGARCH-Copula模型的投资者情绪与股市收益动态相关结构分析

Analysis of Dynamic Correlation Structure Between Investor Sentiment and Stock Market Returns Based on Time-varying ARMA-EGARCH-Copula Model

基于上证市场2011年7月~2017年3月数据,选取流通市值加权换手率、市场融资余额变化率、市场融券余额变化率、涨跌比、基金指数收益率等五项指标对投资者情绪进行测度,并通过时变ARMA-EGARCH-Copula模型实证分析了投资者情绪与市场收益率之间的动态相关关系。分析结果表明,样本期内,上证指数收益率和投资者情绪都存在明显的波动集聚性和持续性;投资者情绪与指数收益率之间总体正相关,但时变性较大。总体上,市场正常波动时两者相关系数与指数收益率大小呈负相关关系,而市场异常波动时,暴跌阶段负相关性有所强化,暴涨阶段负相关性有所弱化。本研究有助于金融监管部门适当引导投资者行为,以减少市场非正常波动,促进金融市场正常发展。

Based on the data of Shanghai stock market in July 2011 to March 2017.Five indicators to measure investor sentiment were selected to measure investor sentiment.These indicators include turnover rates,change rate of the financing balance,change rate of the margin balance,ADR and fund index yield.The dynamic correlation structure between investor sentiment and market returns could be analyzed by time-varying ARMA-EGARCH-Copula model.It was showed that during the sample period,the Shanghai stock index returns and investor sentiment have obvious volatility clustering and persistence;investor sentiment is positively correlated with the index yield,but it obviously exists large time variability.Generally,when the market fluctuates normally,the correlation coefficient is negatively correlated with the index yield.But when the market fluctuates abnormally,the negative correlation is strengthened in the crash phase,and the negative correlation is weakened in the rising phase.This research would help the financial regulators to guide investors＇ behavior properly in order to reduce the abnormal fluctuation of the market and promote the normal development of the financial market.