预期信念中含一般函数参数的资产定价模型及实证研究

发布时间：2018-03-25 16:07

本文选题：函数参数　切入点：历史信息记忆参数　出处：《新疆大学》2017年硕士论文

【摘要】：在经典的经济资产定价模型理论中,假定的是基本面分析者预期信念中价格在一定时间会偏离长期基准价格但最终会向基准价格回归,而仅考虑方差是一个常数,在本文中基本面交易者的价格波动不仅受到当前价格自身的影响,还会受到当前价格和基准价格偏差的影响,而图表分析者相信未来价格的预测来自于当前价格和历史价格的学习过程,这个历史价格过程是一个有限的几何衰减过程,选择的历史信息记忆参数为一个常数,因此,在本文中,我们预期信念中记忆参数选择为一个一般函数,由此构建了一个预期信念中有一般函数参数的资产定价模型.首先,利用差分方程相关理论分情况的讨论了确定性模型平衡解的存在性和局部渐进稳定性,函数参数在满足连续可导的条件下一般平衡解的分支稳定情况,最后,将一般函数转化为一个二次多项式,用模拟和实证研究,通过将本文所建立的模型与原文当中所建立模型数据的统计特征作对比,参照真实市场数据收益序列的一些统计分析,可以发现本文中的模型能更好的反映真实市场特征。
[Abstract]:In the classical economic asset pricing model theory, it is assumed that the fundamental analysts expect that the price will deviate from the long-term benchmark price in a certain period of time but will eventually return to the benchmark price, but only considering the variance is a constant. In this paper, the price fluctuations of the fundamental traders are affected not only by the current price itself, but also by the deviation between the current price and the benchmark price. The chart analysts believe that the prediction of future prices comes from the learning process of current and historical prices, which is a finite geometric decay process, and the historical information memory parameters selected are a constant. In this paper, we select the memory parameters in the expected belief as a general function, and then construct an asset pricing model with general function parameters in the expected belief. The existence of equilibrium solution and local asymptotic stability of deterministic model are discussed by using the relevant theory of difference equation. The bifurcation stability of general equilibrium solution under the condition that the function parameter satisfies the condition of continuous derivation is discussed. The general function is transformed into a quadratic polynomial. By means of simulation and empirical research, the statistical characteristics of the model established in this paper are compared with that of the model established in the original text, and some statistical analyses of the real market data return series are made. We can find that the model in this paper can better reflect the real market characteristics.
【学位授予单位】：新疆大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O212.1

【相似文献】