Tous les séminaires
Jeudi 20 Octobre
Andre LUCAS (VU University Amsterdam)
"Maximum Likelihood Estimation for Generalized Autoregressive Score Models"
(joint work with Francisco Blaques and Siem Jan)
We establish the strong consistency and asymptotic normality of the maximum likelihood estimator for a class of time-varying parameter models driven by the score of the predictive likelihood function; see Creal et al. (2011,2013) and Harvey (2013). Score-driven models possess information theoretic optimality properties, are straightforward to estimate, and easy to generalize to new time-varying parameter settings. We formulate primitive (rather than high-level) conditions for identification, invertibility, strong consistency, and asymptotic normality under both correct specification and mis-specification of the model. Our results are of a global rather than a local nature. We illustrate how to apply the theory using a score-driven time-varying parameter model under skewness and fat tails.
(joint work with Francisco Blasques and Siem Jan)
Prochaine séance : Jeudi 24 Novembre
- Jean-Paul RENNE (University of Lausanne)
- Sébastien FRIES (CREST)