Time Series of Count Data: Modelling and Estimation

This paper compares various models for time series of counts which can account for discreetness, overdispersion and serial correlation. Besides observation- and parameter-driven models based upon corresponding conditional Poisson distributions, we also consider a dynamic ordered probit model as a flexible specification to capture the salient features of time series of counts. For all models, we present appropriate efficient estimation procedures. For parameter-driven specifications this requires Monte Carlo procedures like simulated Maximum likelihood or Markov Chain Monte-Carlo. The methods including corresponding diagnostic tests are illustrated with data on daily admissions for asthma to a single hospital.