基于离散选择模型的内幕交易甄别研究

发布时间：2018-07-17 17:23

【摘要】：内幕交易是证券市场上基于信息不对称的一种不道德行为,给投资者和市场带来危害,并有损股票市场“公开、公平、公正”原则,被各国监管机构所禁止。尽管各国一直致力如何防止内幕交易的发生,但内幕交易案件仍屡禁不止,加之内幕交易行为的复杂性,使得内幕交易的监管执法难以有效实施,因此怎样才能有效的控制内幕交易的发生,一直是各国学者研究的重点。针对这一难题,拟试着研究内幕交易的甄别模型体系,通过内幕交易发生概率的预测,对有可能发生内幕交易的股票进行及时监管和处罚。首先采用均值相等的T检验和非参数检验,筛选出各类指标变量,然后分别建立基于单一指标类的第一阶段离散选择模型。最后按照两种思路建立第二阶段多因素的Logistic和Probit甄别模型,以提高模型的甄别效率,为我国股票市场对内幕交易的监管提供一定的依据。实证结果表明：基于离散选择方法得到的甄别模型,各指标相同且符号相同,并与预期相同；甄别模型与监管类指标、公司治理类指标、市场类指标联系紧密,与财务指标关系不显著,多因素甄别模型以股权集中度、日均异常收益率、跌停次数为自变量,模型的甄别效率达到90%；在甄别效率上,Logistic和Probit模型在各模型中的甄别效率基本上相同；第二阶段的模型甄别效率明显高于基于单一指标类的甄别模型；通过控制显著性水平的变化,监管机构可以达到最优监管的目的,有利于监管部门“相机抉择”；监管部门可以根据各股票内幕交易发生的概率,对股票进行颜色预警。
[Abstract]:Insider trading is an immoral behavior based on asymmetric information in the securities market, which brings harm to investors and the market, and damages the principle of "openness, fairness and justice" in stock market, which is prohibited by the regulatory bodies of various countries. Despite the efforts of various countries to prevent insider trading from happening, insider trading cases are still repeatedly prohibited, and the complexity of insider trading behavior makes it difficult to effectively enforce the supervision and enforcement of insider trading. Therefore, how to effectively control the occurrence of insider trading has been the focus of scholars all over the world. In view of this problem, this paper tries to study the discrimination model system of insider trading, and through the prediction of the probability of insider trading, we can supervise and punish the stocks that may have insider trading in time. First, T-test and non-parametric test with equal mean value are used to screen out all kinds of index variables, and then the first-stage discrete selection model based on single index class is established respectively. Finally, the Logistic and probit discriminant models of the second stage are established according to the two ideas, in order to improve the screening efficiency of the model and to provide some basis for the supervision of insider trading in China's stock market. The empirical results show that the discriminant model based on discrete selection method has the same indicators, same symbols and the same as expected; the discriminant model is closely related to the regulatory index, corporate governance index, market index, and so on. The relationship with financial indicators is not significant, multi-factor screening model to equity concentration, abnormal rate of return per day, the number of times to fall as independent variables, the screening efficiency of the model reached 90. The discrimination efficiency of Logistic model and probit model is basically the same in each model, the second stage model is obviously higher than that based on a single indicator class, and by controlling the change of significance level, the discrimination efficiency of Logistic model and probit model in each model is basically the same, and the discrimination efficiency of the second stage model is obviously higher than that of the model based on single indicator class. Regulators can achieve the goal of optimal supervision, which is conducive to the "discretion" of regulators; regulators can give color warning to stocks based on the probability of insider trading in each stock.
【学位授予单位】：青岛大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F830.91;F224

【参考文献】