This thesis explores the impact of flash crashes on the dynamics of financial markets with asymmetric information. We built, implemented, and analysed an agent-based model of an extended information-sequential trading framework inspired by the models of Das and Glosten-Milgrom, where an exogenous fake shock is added into the system to disturb the actions of some traders where there is informational asymmetry. The key modelled agents include fundamental traders, who place orders at preferred prices; zero-intelligence traders, who place orders randomly; a market maker, who provides liquidity; and an exchange matching all orders under continuous auctions or batch auctions. To this end, by Monte-Carlo methods, we implement the model and examine the dynamics of the market under information asymmetry in the following aspects: the market structure, market risk, the network topology of agents and market mechanisms. Our results demonstrate that, an uninformed fundamental trader (UFT) in a messy network is highly likely to suffer a major loss due to the significant price crash in a strongly UFT-dominated market (the informed traders only account for less than 20%), in which case the market efficiency is also negatively affected; Applying batch auctions helps reallocate the profits among the agents to reduce the information advantage between informed and uninformed traders, but it has limited effect on mitigating flash crashes; Building an information-sharing connection between agents is effective to reducing flash crashes and narrows the information advantage gap between informed and uninformed traders, but a complete network with full information exposure could mislead uninformed traders to make biased decisions. These findings generated by an agent-based simulation model give us insights into real-world financial markets under asymmetric information, and the framework proposed in this thesis can be extended for future studies of asymmetric-information markets.
Understanding How the Flash Clashes are Affected in an Asymmetric Informational Market with Agent-based Modelling
Sun, Y. (Author). 1 Dec 2023
Student thesis: Doctoral Thesis › Doctor of Philosophy