Title: An information lag component in spread decomposed model
Authors: Qiang Zhang - Beijing Universtiy of Chemical Technology (China) [presenting]
Abstract: One order Markov property of trade indicator variables as a key assumption in MRR model contradicts with information lag as an empirical characteristic in high frequency trading process. In this paper, a nonparametrical test is employed and the Markov property of trade indicator variables is rejected in most trading days. Based on the spread decomposed structure of MRR model, a moving average structure is adopted to absorb the information lag as an extension, then a ML estimator with ARCH structure is introduced. Empirical results show that the information lag parameter is significant and the adverse selection risk parameters estimated by the original and the extended, respectively, have significant differences.Further, the analysis suggests that the information lag parameter could measure the average speed at which the information incorporates into the price.