A1551
Title: Modelling dynamic multiple quantiles with exogenous factors
Authors: Pierluigi Vallarino - Aarhus BSS (Denmark) [presenting]
Alessandra Luati - Imperial College London (United Kingdom)
Leopoldo Catania - Aarhus BBS (Denmark)
Abstract: A new dynamic model for the quantiles of the conditional distribution of a random variable is introduced. The model incorporates exogenous covariates by writing the quantiles of interest as the sum of two quantile functions: the first one subsumes past information on the variable of interest; the second function is a linear combination of baseline quantile functions, each one related to one of the predictors. The dynamics of the quantile functions rely on a set of martingale differences sequences arising from a quasi-score-driven approach. Parameters of the model are estimated through a two-stage quasi maximum likelihood estimator (2SQMLE). Consistency and asymptotic normality of the 2SQMLE are derived, and its finite sample properties are assessed through a simulation study. An empirical analysis concerning macro-financial variables shows that the default spread is relevant for predicting the left tail of the conditional distribution of the industrial production index.
