Empirical study of a bid-ask model for liquid markets
Douglas Machado Vieira
market microstructure | limit order book | empirical analysis | queueing models
The present thesis is devoted to empirically study the Markovian micro structure model investigated in Cont and de Larrard (2013), which is intended for liquid markets. Besides other results, the model is able to provide distributions for the durations between mid price changes, probabilities of mid price changes conditioned to the state of the order book and, more surprisingly, link pure micro structure statistics with volatility. With the aid of a freely available high-frequency dataset provided by NYSE, we extensively investigate the assumptions of the model and some of its results for the stocks from the Dow Jones Industrial Average Index, one of the most liquid markets in the world. In this thesis, we conclude that, although there are various unrealistic assumptions, the model is still able to retain the core of the high frequency mechanics and produce consistent results, including the aforementioned volatility relationship.