Abstract: Our sample include all NYSE, Amex and NASDAQ common stock for the period from July 1962 through December 2014, using midpoint of bid-ask quotes, bid price and return weighted these three ways to eliminate the effects of prices' noise in stock returns and stock portfolio returns, reexamining MAX effect. Meanwhile, considering the fact that more than 53% in the extreme daily returns stocks belongs to NASDAQ, this paper also test three exchange markets separately. The empirical results do not give evidence that interpret the "MAX anomaly", suggests that combination of market microstructure noise and Fama-French-Carhart four factor model, as well as market characteristics, is not the causes of the "MAX anomaly".

Key words: MAX anomaly, microstructure noise, four factor model, security exchange market