状态依赖下的股市正反馈交易Positivefeedbacktradinginequity.PDF

状态依赖下的股市正反馈交易 方毅 1, 2，张屹山 1, 2，张代强2 (1．吉林大学数量经济研究中心，长春 130012； 2．吉林大学商学院，长春 130012) 摘要：在 Sentana 和 Wadhwani 的正反馈模型基础上，本文利用前期涨跌进行状态划分，建立了正反馈交易 的 Markov 状态依赖模型，考察市场中的正反馈交易行为与羊群效应。通过对中国股市的实证研究，可以 发现，股价不服从随机游走，在前期股价连续上涨或下跌的情况下，转移概率具有惯性；中国股市存在显 著的状态相依的正反馈行为，在前期股价连续上涨的情况下，易激发羊群行为。监管者应注重市场价格的 持续变动，投资者可以采用正反馈交易策略。 关键词：状态依赖；正反馈；羊群效应；转移概率 中图分类号：F830.91 文献标识码：A Positive feedback trading in equity market depends on states FANG Yi1，2 Zhang Yishan1，2 Zhang Daiqiang2 (1. Center for Quantitative Economics, Jilin University, Changchun, 130012, China; 2. Business school, Jilin University, Changchun, 130012, China) Abstract: Based on the positive feedback model of Sentana and Wadhwani, this paper establishes the Markov state dependent model with the positive feedback trading, and studies the positive feedback trading behavior and the herd effect in the market. According to the empirical study in the Chinese equity market, it can be found that the stock price doesn’t obey the random walk after prior continuous rising. The transition probability has the inertia in the condition of the prior continuous rising or dropping. Moreover, the Chinese equity market has the significant positive feedback behavior depending on states. The herd behavior can be easily simulated when the price continuous rising. The regulator should pay attention to continuous change of the market price. In the mean time, the investor may take advantage of the positive feedback trading strategy. Key words: state dependent; positive feedback; herd effect; transition probability 课题资助：本文受国家自然科学基金项目“基于证券市场效率的投资者总体偏好非参数检验” （项目号）资助，吉林大学基本科研业务费项目“股票市场羊群行为的非线性特 征研究”（项目号：2011QY094 ）资助。