Model order selection for cascade autoregressive (CAR) models

Loading...
Thumbnail Image

Date

2021

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

In recent years, Cascade Autoregression (CAR) models enjoy increasing popularity in applied econometrics. This is due to the fact that they are able to approximate both short- and long-memory processes and are easy to implement. However, their model order, namely the timing of the steps, relies on ad-hoc decisions rather than being data-driven. In this paper, techniques for model order selection of CAR models in finite samples are presented. The approaches are evaluated in an extensive simulation study, as well as in an empirical application. The results suggest that model order selection may provide gains in both in- and out-of-sample performance.

Description

Table of contents

Keywords

cascade models, cross-validation, forecasting, time-series modeling, long-memory, model selection, HAR

Citation