A Study On Information Value Of Earning Preannouncements Of Listed Companies And Trading Strategy

It is required by the Chinese securities regulatory authorities that listed companies should publish earning preannouncement before the announcement of formal periodical financial reports if loss or substantial performance changes are expected during the report period. Based on the total sample, this dissertation makes comparison test between the before and after event windows to explore the information value of earning preannouncement. What’s more, the author develops corresponding trading strategy based on the change law of share price’s statistic characteristics, and finally proves again the information value of earning preannouncement on guiding trading under the systematic evaluation of trading strategy.The research finds that stock price returns do not show significant change while the return’s standard deviation, that is volatility, rises thirty days before and after the announcement day. To explain it, this dissertation analyses this issue from the perspectives of the behavioral financial theory and stock operation techniques. It is found that several essential factors lead to this issue:a dynamic process from under reaction to overreaction and finally to equilibrium, which is brought by investors’ heterogeneity; a market manipulation action drove by the main funding using their information advantages.Inspired by the findings above, the author develops the moving average trading strategy based on HP filtering and the adaptive fractal trading strategy. The former strategy increases the signal accuracy on the basis of the traditional trading strategy that signals are produced by the crossing of moving average, and the noise elimination filters additionally, producing good results. However, further analysis shows that great subjectivity exists in parameter selection and threshold setting, thus a new adaptive selection method should be considered. The adaptive fractal strategy develops from a complete classification of K-line positional relations and gives a way like this. And the introspection results indicate that the adaptive fractal trading strategy is more beneficial and has less withdrawal possibility.At last, the White Reality Check and random sample test are adopted to test robustness of adaptive fractal trade strategy. Results demonstrate the great robustness exist under different situations, ruling out the possibility of overidentification which would lead to one sample with good performance by chance. Moreover, the main theme of this dissertation is verified again--earning preannouncements do have significant information value.