Title: Market Microstructure: from equilibria to VAR models

Abstract:

Market microstructure as a branch of financial economics plays a crucial role in studying the price formation process in the security market. In this dissertation, we build an integrated market microstructure framework called the Bi-Noise model that bridges game-theoretic modeling and econometric analysis. In the first part, we formulate an extensive-form game within three distinctive participants and rigorously present the explicit form of the equilibrium price by accommodating the price impacts from three key market frictions: inventory control, adverse selection, and rationally uninformed trades. An empirical counterpart of the Bi-Noise model is also proposed, and with Bayesian MCMC algorithms utilized, numerical experiments are conducted in terms of both the synthetic data and the real trading data. In the second part, we use the Bi-Noise model to deliver an empirical analysis according to each type of price impact in a temporal manner. By considering the expression for the quote update, we also show that the practical Bi-Noise model can be transformed into a vector autoregression (VAR) time series framework. Finally, we put forth an alternative parameter estimation approach with a hybrid inference structure and apply it to the empirical study.

Share this Event