Sources Of Industry Momentum Profit, Autocorrelation Of Returns And Lead-lag Relationship

In this paper,the identity decomposition method proposed by Lo and Mackinlay is used to decompose momentum profit into cross-autocovariances of industry portfolio returns,own-autocovariances of industry portfolio returns,and the cross-sectional dispersion in mean portfolio returns,in order to conduct an empirical study on the profit sources of the industry momentum effect of China's A-share market.Compared with previous studies,the identity decomposition method proposed by Lo and Mackinlay in this paper can obtain higher-order cross-autocovariances of industry portfolio returns and own-autocovariances of industry portfolio returnsThe sample data in this paper is the weekly returns of 28 shenwan industry portfolios from 2006 to 2019.The industry momentum profit decomposition result shows that the source of industry momentum profit in China's A-share market is mainly the own-autocovariances of industry portfolio returns that are significantly positive.And cross-autocovariances of industry portfolio returns weakens momentum trading profits,the cross-sectional dispersion in mean portfolio returns has little effect on the profit of the industry's momentum strategy.Therefore,the profit of the industry momentum strategy in China's A-share market comes from the autocorrelation of the industry portfolio return,which is in line with the BSV model,DHS model,and HS model proposed by behavioral finance.At the same time,the industry momentum profit gradually increased during the lag 1-3,and the own-autocovariance of industry portfolio returns changed little.The increase in the industry momentum profit is due to the cross-autocovariances of industry portfolio returns.The increase in industry momentum profit is due to the gradual decrease in cross-autocovariances of industry portfolio returns.This article further analyzes the sources of autocorrelation of industry portfolio returns.The autocorrelation of industry portfolios may come from asynchronous transactions and lead-lag effects.This article analyzes the momentum momentum profit of the low-frequency 2-week data and 4-week data.The results show that after excluding asynchronous transactions,the own-autocovariances of industry portfolio returns even rises slightly,so asynchronous transactions cannot explain the autocorrelation of industry portfolio returns.Secondly,this paper builds alphabet-based portfolios to study the impact of the lead-lag effects.The empirical results show that the short-term momentum profit of alphabet-based portfolios also comes from the ownautocovariances of the portfolio returns,but its value is smaller than the ownautocovariances of industry portfolio returns.Therefore,the lead-lag effect of stocks in the portfolio can only partially explain the source of autocorrelation of portfolio returns.