反馈交易规则与股票收益自相关实证分析

研究了反馈交易规则与上证综合指数日收益自相关之间的联系，样本取自上证指数１９９２年９月２９日至１９９７年９月３０日的日收益率数据。此外，文中采用ＧＡＲＣＨ（１，１）处理收益波动的异方差性，模型参数采用极大似然估计。实证模型还考虑了非同步交易引起的工序列相关以及反馈交易的非对称性。结果表明，除以往文献涉及的正序列相关外，正反馈交易将导致收益负的序列自相关，且相关系数绝对值随波动增大而增大，从而整个收益表现出随波动变化的序列相关。

Empirical Evidence of Feedback Trading and Stock Return Autocorrelations

This paper studies the pattern of autocorrelation of stock returns in Shanghai stock market, assuming that some investors follow a positive feedback trading strategy. The sample is the daily returns from September 29 of 1992 to September 30 of 1997. In addition, GARCH(1,1) model is adopted to deal with heteroskedastic property of return volatility, and the maximum likelihood method is used to estimate the model parameters. The positive serial correlation caused by nonsynchronous trading and nonsymmetry of feedback trading are also considered in the model. The empirical results show that in addition to positive serial correlation described in previous literature, there is a strong evidence that positice feedback trading induces negative autocrrelation in index stock returns, which in absolute terms increases with volatility. Thus the autocorrelation of index stock returns and volatility are related.