First, theoretical analysis is made on the difference and the adaptability of three models of measuring long-term abnormal returns, then simulation is employed to detect the utility of these models and the conclusion that TBHAR is better than the other two is made.

Conducting with the method of event study, this article chooses randomly about 200 stocks as samples in Shanghai A share market and calculates the cumulated abnormal return of noise traders by constructing the markup portfolios of positive feedback trading.

The study finds that market response positively to the release of the reform information,cumulative abnormal return is positive,and the extent of response is significantly positive with the ratio of presenting shares and insignificant with the proportion of non-circulate shares.

The article analyses the method used by Chinese researcher in calculating and testing long-run abnormal returns(calculating long-run abnormal returns by market portfolio,testing by standard t test),point out its deficiency,gives new method to overcome it.