Ph.D. Defense: Huijun Qian
Ph.D. Thesis Defense
Friday, April 5th, 2019
112 Hanes Hall
Statistical Inference and Computation for Market Microstructure
(Under the direction of Dr. Chuanshu Ji)
Our work in this thesis focuses on studying three main aspects of market microstructure that attract most attentions from financial economists: asymmetric information, bid-ask spreads, and the impact of trading mechanisms on price formation. To study these, two model-based approaches are adopted here: an extended G-M and a modification of Hasbrouck model. The extended G-M is a sequential trade model assuming risk neutrality and a quote-driven protocol, conditioning on the underlying value of the asset, it can present explicitly how bid and ask prices change over time and are influenced by different trading orders. The dynamic trading flow with three trading options (buy, sell or no trade) and corresponding volumes are modeled by adopting the stochastic supply curve. The modified Hasbrouck model adopt bivariate time series to model the relationship between trading volumes and the midpoints of the bid-ask spread intervals. Such bi-linearity reduces the computational burden. For both models, two different volatility settings: constant and GARCH(1,1) are considered and Bayesian MCMC algorithms are applied for both the simulation study and empirical study.