基于计算金融实验的市场崩溃与模仿式羊群行为研究

Stock market crashes, one kind of excess volatility, are momentous financial events that are fascinating to academics and practitioners alike. According to the academic world view that markets are efficient, only the revelation of a dramatic piece of information can cause a crash, yet in reality even the most thorough post-mortem analyses are typically inconclusive as to what this piece of information might have been. For the heterogeneous traders and investors, the fear of a crash is a perpetual source of stress, and the onset of the event itself always ruins the lives of some of them..The finding of excess volatility points to intrinsic dynamic forces of speculative markets not related to fundamental factors. It has been considered as "non-standard" models focusing on fads and sociological or psychological mechanisms. A related empirical finding is the recent evidence on mean-reversion in asset prices. Technically, this means that there is positive autocorrelation over longer intervals in the data. Models, called herd behavior (HB) in all, are developed to understand this regularity. They showed the importance of psychological factors and irrational factors in explaining historical financial crises. But first, the descriptive interprets of HB can not uncover the potential relating mechanism between the market crash and the heterogeneous participators. And second, indeed, the trade tactics of HB are private and recessive. In fact these models have deficiencies. Following these reasons, the aim on answering these questions is to construct an elementary model of stock market dynamics which explicitly includes contagion of opinion and behavior and to offer a behavioral explanation for the empirical findings discussed above. But how can we deal with such complex finance system?.Considering this situation, the project studing HB in a laboratory financial market. And the purposes are: ①investigating HB's prewarning effect on the market crash, and disclosing the potential relating mechanism between the market crash and the heterogeneous participators; ②studing the connecting link between the market's macro scenes and the individual's micro actions; ③building the common HB model based on imitation; ④developing the imitate HB model to the additional market's public information; ⑤constructing the forecast policy research of the market's crash. .In brief, the market of "Chinese background" structures that these researches going to uncover in this project may be understood as the robust emergent properties of the market resulting from some characteristic "rules" of interaction between investors. And it carries the prospect of promoting the creatively and deeply native research of behavioral finance. Further task is to guide the decision of individual investors, and to shred light on the way of regulator's policies of the certain financial market.

金融市场崩溃与羊群行为存在事实上的关联机制：发生前市场常有超过1年的持续的、逐渐加快的上涨；崩溃发生时多令投资者不及反应，及伴有显著的投机、泡沫、市场恐慌等一致性特征。市场崩溃广泛频发，破坏力大，造成影响广，持续时间长，从理论研究和实践者及监管方看来都显得尤为迫切和极端重要，是值得探究的重大金融问题。因此，通过深入探讨羊群行为来了解市场崩溃的形成机制与警示信号的研究也十分必要。考虑到羊群行为的基本特征和信息私密性，本项目借助计算金融实验方法研究：①探求以羊群行为为基础的市场崩溃预警信号，发现市场崩溃的内部机制；②揭示其中的宏观市场现象与微观个体行为的联系纽带；③构建基于模仿的一般的羊群行为模型；④构建考虑市场情绪影响的模仿式羊群行为模型；⑤尝试对市场崩溃进行对策研究。本研究项目拟推动具有本土特点的行为金融研究的创新与深入，从而指导投资者决策和为有关金融政策推出及市场监管提供实验及实证借鉴。

根据课题组的研究，羊群行为作为资本市场异质性特征的集中表现方式，也是本课题的计算实验需要跟踪的的Agent的关键行为，其在新兴市场的表现相较于传统市场更为显著。这是因为新兴市场的体量、监管以及信息消化效率都要弱于传统市场，这些特征使得新兴市场有很大的可能成为非有效或部分有效的市场。这样一来，发掘羊群行为这一市场异象必须关注新兴市场的计量特征，从而也为市场间异质性提供了研究支撑。这也是本报告的第一部分的主体内容。提炼市场结构的计量特征是构建Agent生存环境的基本参考。课题组经过充分研究与谈论，确定将市场流动性作为仿真平台市场结构构建的基本参考。因为Amihud曾指出：“流动性是市场的一切”。在对流动性的研究上，课题组将研究重点放在高频数据流动性的研究上。高频数据的日内特征中富含了丰富的流动性信息，而这些流动性信息中又是对资本市场波动高度关注的群体某种程度上的羊群性的集中体现，因为日内信息中的跟随者对信息交易者的跟踪具有很强的主观依赖性。这方面的研究形成了本报告的第二块主体内容。当课题组将研究重点转向对市场崩溃的发生机制与后续影响进行建模和实验处理时，发现2008年的金融危机就后果来说，其摧毁性影响和传递机制是明确的，但是其与1929年的世界经济大萧条和一百多年前发生在英国、法国以及荷兰的经济危机相比，存在着期限和结构不一致的特征。就期限来说，一百多年前的英国、法国和荷兰的经济危机是资本市场初级阶段的非稳态原发性危机，而1929年的危机是全球化商品市场发展落后于产能的危机，2008年的危机则是全球化房地产市场消化模式不健全的危机。事实上，2008年的金融危机之后，一个不被人注意的后果是，房地产市场的消费模式真正走入了全球化。由此，课题组一致认为，在基于市场崩溃的羊群行为研究中，计算实验方法应重点刻画危机消散的代价：不是市场重建，而是市场全球化的突破性演化机制。因此，结合课题组在第一部分研究工作中所积累的对碳交易这一新兴市场的深入研究与理解，又引入了气候变化和公司治理关系作为碳市场交易的前提和度量手段，从而得以探讨在计算实验环境下验证碳市场崩溃所引发的全球市场动荡与碳市场的全球化发展预测。这三方面的研究都构成了对市场崩溃与模仿式羊群行为的创新性的理解与判断。课题组已开始关于碳市场的全球化特征的思考与探索，并结合对羊群行为的基本特征的总结与整理及其计算实验.