High frequency stock returns may show mild positive autocorrelation because of non-trading. However in a noise traders model the degree of autocorrelation may change with volatility: if some traders follow feedback strategies, variations in volatility affect the degree to which feedback traders influence the share price. Further, if risk aversion declines with wealth, the extent of positive feedback trading may increase when volatility increases.
The results reported in this paper seem to be in accordance with this view: when the volatility is low, daily (and hourly) stock returns exhibit positive serial correlation. However, when volatility is high, returns exhibit negative serial correlation. Our results also suggest an important asymmetry - returns are more likely to exhibit negative serial correlation after price declines. This is consistent with prices declines being more likely to induce positive feedback trading. We also find no significant relation between margin requirements and the serial correlation of returns.
Download is not available
