earticle

영어

This paper investigates possible explanations for the customer momentum documented by Cohen and Frazzini (2008). First, returns to the customer momentum vary significantly over time and can have large crashes during market crises. However, I do not find evidence linking the customer momentum to macroeconomic risk. Second, the customer momentum does not weaken significantly out-of-sample and thus is unlikely to be driven by statistical bias. Third, there is no evidence for the publication effect of Mclean and Pontiff (2016). Hence, the customer momentum does not represent profitable opportunities for arbitrage. Consistently, I show that the customer momentum is concentrated in stocks that are difficult to arbitrage. Moreover, trading on the customer momentum generates high turnovers but the probability of underperformance is not as high in annual frequency. Overall, my evidence suggests that limits-to-arbitrage prevent the incorporation of new information on customer firms into the share prices of suppliers.