中国上市公司盈余预告公告的信息价值及交易策略研究

发布时间：2018-03-07 15:14

本文选题：盈余预告　切入点：信息价值　出处：《厦门大学》2014年硕士论文　论文类型：学位论文

【摘要】：中国的证券监管机构规定,上市公司在正式的定期财务报告公布前,如果预计报告期出现亏损或大幅业绩变动,必须发布盈余预告公告。本文的研究旨在从全样本的角度出发,借助前后事件窗口的比较检验,探究预告公告的信息价值,并根据预告前后股价统计特性的变化规律开发出相应的交易策略,最后通过对交易策略的系统性评价再次论证盈余预告公告所具有的能够指导交易的信息价值。研究发现,在公告日前后三十天的时间窗口中,股价收益率并没有明显的改变,而其收益率的标准差即波动性却变大了。对此本文从行为金融理论和业界操盘手法两个角度进行了分析,认为造成上述结论的直接原因是在公告日后出现了先涨后跌或先跌后涨的现象,而这种现象的本质成因是有多种可能的,可能是投资者异质性交易带来的对新信息从反应不足到反应过度到均值回复的动态过程,也可能是主力资金根据其信息优势推动的一种市场操纵行为。基于上述发现,本文开发出了基于HP滤波的MA均线交易策略和自适应分型交易策略。前者主要是在传统的移动均线交叉产生信号的策略基础上,辅以滤波去噪来提高信号准确率,取得了不错的绩效,但进一步的分析表明,参数选择和阈值设定具有较大的主观性,因此考虑一种新的自适应选择的方法。自适应分型即从K线位置关系的完全分类出发,给出了这样一种方式。回测结果表明,自适应分型策略具有更好的收益以及更小的回撤。最后,本文采用White真实性检验和随机样本检验的方法对自适应分型策略进行了稳定性测试,结果表明其具有不同样本下良好的稳定性,并非过度识别导致碰运气地选到一个刚好有良好绩效的样本,从而再一次地论证了本文的一大主题,盈余预告公告信息具有显著的信息价值。
[Abstract]:China's securities regulator stipulates that if a listed company is expected to report a loss or a substantial change in its performance in the reporting period before the publication of its formal regular financial report, it must issue a notice of earnings. The purpose of this study is to start from the perspective of a full sample. With the help of the comparative test of the event window before and after, the information value of the notice announcement is explored, and the corresponding trading strategy is developed according to the changing law of the statistical characteristics of the stock price before and after the forecast. Finally, through the systematic evaluation of trading strategy, the paper demonstrates the information value of earnings announcement. The study found that there was no significant change in the stock price return in the 30 day window around the announcement date. However, the standard deviation of the return rate, that is, the volatility, has become larger. It is considered that the direct cause of the above conclusion is the phenomenon of rising first and then falling or falling before rising after the announcement date, and there are many possibilities for the essential causes of this phenomenon. It may be the dynamic process of investors' heterogeneous trading from underreacting to over-reacting to average recovery, or it may be a kind of market manipulation driven by the main fund according to its information advantage. Based on the above findings, the MA mean line trading strategy based on HP filter and the adaptive classification trading strategy are developed in this paper. The former is mainly based on the traditional mobile mean line cross generation strategy. With the help of filtering and denoising to improve the signal accuracy, good performance has been achieved, but further analysis shows that parameter selection and threshold setting are more subjective. Therefore, a new method of adaptive selection is considered. The adaptive classification is based on the complete classification of K-line position relationship. The results of the back test show that the adaptive classification strategy has better revenue and smaller withdrawal. Finally, the White authenticity test and random sample test are used to test the stability of the adaptive classification strategy. The results show that it has good stability under different samples. It is not excessive recognition that leads to the lucky selection of a sample with good performance, which proves once again that the earnings announcement information has significant information value.
【学位授予单位】：厦门大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.51;F231.1

【参考文献】