反馈型交易策略与资产价格演化

摘要

This paper studies how investors' feedback trading strategies and learning mechanism affect asset price dynamics. First, Markov chain method is used to characterize the feedback investors' belief updating mechanism. Second, by using evolutionary game model we analyze the evolution of positive feedback trading strategy and negative feedback trading strategy in the market. Third, assum-ing that the market consists of fundamentalists and feedback investors, we develop a model of risky asset price dynamics, and analyzes how fundamentalists' information collection cost influence their arbitrage behavior. Simulation results show that when fundamentalists' in-formation collection costs greatly, their arbitrage behavior is restricted. In the game with fundamentalists, positive feedback traders sur-vive, and gradually drive fundamentalists and negative feedback traders out of the market. Thus, asset price continues to deviate from the fundamental, and the wealth of positive feedback traders gradually increases whileas the wealth of fundamentalists sharply reduces. Models proposed in this paper provide possible explanations for the causes of bubbles and anti-bubbles, indicating that market supervi-sion departments should try to reduce the investor's information collection costs to enhance market efficiency.%　　本文研究投资者的反馈型交易策略和学习机制是如何影响资产价格演化的。首先，运用Markov链方法刻画了反馈型投资者的信念更新机制；其次，运用演化博弈模型分析了正反馈及负反馈这两种交易策略在市场中的演化；最后，构建了市场由基本面投资者和反馈型投资者构成的风险资产价格演化模型，分析了基本面投资者的信息收集成本对其套利行为的影响。仿真结果表明，当基本面投资者的信息收集成本较大时，其套利行为受到限制，持续预期交易者在与基本面投资者的博弈中生存下来，并逐渐将基本面投资者和反转预期交易者赶出市场，使得资产价格持续偏离基本价值，持续预期交易者的财富逐渐增加，而基本面投资者的财富急剧减少。本文提出的模型为泡沫及反泡沫产生的原因提供了一种可能的解释，同时表明，市场监管部门应尽力减小投资者的信息收集成本，从而增强市场的有效性。