CAJ | 학술논문

资产收益存在长期相关性是金融市场的一个普遍现象。提出一个交易者具有不同预期信念的资产定价模型，试图解释收益的尖峰胖尾、波动聚集以及长期记忆性等格式化特征．引入一种记忆机制解释长期记忆产生的原因。模型中交易者被分为基础交易者和图表交易者，他们对价格的均值和方差具有不同的信念，且根据交易策略的适应度来选择交易策略，交易者之间互动以及与做市商的相互作用驱动资产价格做非线性的运动。最后对模型数据进行R／S分析，并使用ARFIMA和FIGARCH模型拟合，确认了数据中长期记忆性的存在。
자산수익존재장기상관성시금융시장적일개보편현상。제출일개교역자구유불동예기신념적자산정개모형，시도해석수익적첨봉반미、파동취집이급장기기억성등격식화특정．인입일충기억궤제해석장기기억산생적원인。모형중교역자피분위기출교역자화도표교역자，타문대개격적균치화방차구유불동적신념，차근거교역책략적괄응도래선택교역책략，교역자지간호동이급여주시상적상호작용구동자산개격주비선성적운동。최후대모형수거진행R／S분석，병사용ARFIMA화FIGARCH모형의합，학인료수거중장기기억성적존재。
A common feaure found in fincial market is long memory presented in the asset returns. This paper proposes a simple asset pricing model with heterogeneous belief to explain the some stylized facts founed in the distribution of market returns, such as fat-tailness, volatility clustering and long memory. In our model, traders are classified into fundamentalists and chartists who have different expectation about the mean and variance of prices. They choose trading strategies based on their performances. The price fluctuations are driven by the interactions between the traders and market maker. A memory mechanism is also introduced to generate and explain long memory. Finally, the paper confirm the presence of long memory in data generated by our model through R/S analysis as well as esatimations of ARFIMA and FIGARCH model.