Masters in Statistics Oral Defense: "Statistical Analysis of Markovian Queueing Models of Limit Order Books"

Abstract: The objective of this thesis is to investigate the suitability of some Markovian queueing models in being able to effectively describe the dynamical properties of a limit order book more specifically. We review and compare the assumptions proposed by Huang et al.[Quantitative Finance,$\bm{12}$,547-557(2012)] and Cont et al.[SIAM Journal for Financial Mathematics,$\bm{4}$,1-25(2013)], and estimate the intensity parameters in both ways, based on real data of a stock on the Nasdaq Stock Market. Trough comparing by cumulative distribution functions of first-passage time to state 0, we will hsow that the estimators of Cont's model fit our data better and we put forward the assumption of multiple-size rates as a better alternative to Cont's frame work. At last, we investigate the stationary joint distribution of volumes on either side after each price change.

Host: Jose Figueroa-Lopez