Better predictability of arbitrage strategies trained among less attentionstocks

Jie Fang, Jianwu Lin

Abstract: Since less attention stocks have slower information exchange speed, the best parameter set of an arbitrage strategy trained in sample have better predictability out of sample among them than ones among attention stocks. We use two investor-attention sentiment factors to classify the less attention stocks and test both basic momentum strategy and reverse strategy on A-share market over less attention and attention stocks.