Title: Modelling long memory and structural breaks in count data
Authors: Mawuli Segnon - University of Münster (Germany) [presenting]
Abstract: An integer valued negative binomial Markov switching Multifractal (NegBin-MSM) model is developed by adapting the MSM process for count data setting. We provide the statistical properties of the NegBin-MSM process and demonstrate its capacity to reproduce overdispersion, long memory and structural breaks that characterize count data. We show via Monte Carlo simulation that the maximum likelihood estimator is consistent and asymptotically efficient. An empirical application with financial transaction data illustrates the practical importance of the model.