前景理论下投资者风险偏好，决策过程及资产定价

In the last several decades, a complete framework for the academics of asset pricing is built based on the traditional utility function and rational expectation equilibrium. However, the existence of several pricing anomalies, which are inconsistent with the pricing model, implies that the basic hypothesis under the framework could be unrealistic. Meanwhile, the outstanding explanatory ability of Prospect Theory and Mental Account Theory on the studies of decision-making academics receive a mass of attention. The reference-dependent utility function and editing-related mental accounts are considered as one of most promising mechanism for explaining the anomalies relating to the auto-correlation of asset price and price bubble. .In this project, empirical and theoretical approaches will be applied to investigate the roles of previous realized and unrealized profits on both the risk attitude of investors in micro-level and the asset pricing dynamics in macro-level. Firstly, several empirical tests are employed for understanding how the risk-preference of investors varied by the performances of assets held or liquidated by them. Based on the empirical findings, a dynamic model, in which the risk-attitude of representative or heterogeneous agent(s) is impacted by their previous profit, is built. Several possible channels for emerging and vanishing of anomalies are studied based on a nonlinear expectation equilibrium solved by a machine learning approach. Finally, based on the implication of dynamic asset pricing model, this project provides insight for understanding underlying mechanisms of the positive feedback of asset prices and the emerging and vanishing of price bubbles which is the key element in the systematic risk.

传统效用函数与理性预期均衡框架带领着资产定价研究发展至今，已形成完整的理论体系。一些与实证检验结果相违背的异象，意味着该框架存在着与现实违背的假设。与此同时，前景理论及心理账户理论在风险决策研究中的表现在学术界引起广泛关注。基于参考点的效用函数和带有编辑的心理账户是解释收益率自相关，价格泡沫等异象最为可能的机制。.本研究拟采用实证研究和理论建模方法，挖掘交易者实现和未实现的盈利状况如何通过影响微观交易者风险偏好进而影响资产定价宏观涌现。首先通过实证检验方法，挖掘市场参与者的风险偏好以及资产价格是如何受到过去盈利状况的影响。在此基础之上，构建已实现和未实现盈利联合作用的动态资产定价研究框架，并利用机器学习的方法挖掘均衡解，并刻画不同状况下异象的涌现和消失。最后，依据动态均衡模型，尝试挖掘资产价格产生正反馈机制的途径以及系统性泡沫产生和崩溃的机制，为金融系统性风险防范与控制提供新的见解和依据。

投资者偏好与行为是资产价格形成过程中的重要影响因素，其微观的复杂联动可能造成宏观市场的系统性风险。本研究从行为金融学角度出发，挖掘了不同类型的投资者（包含机构和个体投资者）的行为，以及这些行为对微观市场结构和宏观资产定价的影响。具体包括：1）个体投资者的前景理论偏好与短期反转；2）个体反应不足与流动性过度供给；3）基金行为与管理能力评价；4）基于信息不对称的微观市场行为与信息挖掘。本项目通过这些研究拓展了中国证券市场从微观个体行为与市场宏观涌现的相关研究，并为金融系统的市场设计提供了一定的依据。