Morvan Donfack et Arnaud Dufays publient dans Studies in Nonlinear Dynamics & Econometrics

11 septembre 2020

Morvan Nongni Donfack , étudiant au doctorat (Département d'économique de l'Université Laval) et Arnaud Dufays, professeur associé au Département  publient un article intitulé Modeling time-varying parameters using articial neural networks: A GARCH illustration dans la revue Studies in Nonlinear Dynamics & Econometrics. Cet article est issu du du premier chapitre de la thèse de Morvan Donfack.

Ci dessous un résumé du papier.

Abstract : We propose a new volatility process in which parameters vary over time according to an articial neural network (ANN).We prove the process's stationarity as well as the global identication of the parameters. Since ANNs require economic series as input variables, we develop a shrinkage approach to select which explanatory variables are relevant to forecast volatility. Empirically, the proposed model favorably compares with other exible processes in terms of in-sample t on six nancial returns. It also delivers accurate short-term volatility predictions in terms of root mean squared errors and the predictive likelihood criterion. For long-term forecasts, it can be competitive with the Markov-switching GARCH model if appropriate exogenous variables are used. Since our new type of time-varying parameter (TVP) process is based on a universal approximator, the approach can readily revisit and potentially improve many standard TVP applications.