On The Correlation Between Earnings And Stock Returns In China

We focus on the characteristics of the relationship between accounting earnings and stock returns, as well as the characteristics of earnings response coefficient itself in this paper. First, we analyse that earnings response coefficient is nonlinear, different for profits and losses,and different across firms. Then, in summarizing the conclusions of the studies at home and abroad, based on the use of linear model, nonlinear model, as well as the linear and nonlinear model including the firm-specific,we begin the empirical research.We select 2002-2008 Shenzhen A-share listed companies as the research object, respectively, calculating the accumulative abnormal stock return under both the long window period and the short window period,6948 samples are effective in the short window period,6097 samples are effective in the long window period. Under the two window periods, with all samples and profit samples, unexpected earnings and the accumulative abnormal stock return,respectively,have been done a descriptive statistical analysis. Then we compare the earnings response coefficients of linear and nonlinear model of the empirical test, the nonlinear model obtained a better response than the linear model to accounting earnings information. At the same time,in the comparison of all samples and samples of profits under the empirical results, we have the conclusion that the profits samples obtained a surplus of earnings response coefficients than the losses. In the comparison of the regression results of two window period, we found the unexpected earnings is more effective information in a long window period. In the latter part of the empirical content, using a long window period sample data,we analyze non-linear, losses and profits and firm-specific in the relation of returns and earnings,which are expected to help unexpected earnings explain more variation of the accumulative abnormal stock returns. Obtaining that firm-specific gives the largest incremental explanatory power. The non-linear factor and loss-profit factor can not be determined which gives more incremental explanatory power., while one influences another. Taking non-linear problem in the applicability of our country, sample selection and the window period issue into account, the paper provides the improvement suggestion in the future research.