Forecasting Conditional Covariance Matrices in High-Dimensional Time Series: A General Dynamic Factor Approach

36 Pages Posted: 17 Jun 2019 Last revised: 28 Sep 2020

See all articles by Carlos Trucíos

Carlos Trucíos

Universidade Federal do Rio de Janeiro (UFRJ) - Faculdade de Administracao e Ciencias Contabeis - FACC

João Henrique Gonçalves Mazzeu

Universidad Carlos III de Madrid

Marc Hallin

ECARES, Universite Libre de Bruxelles

Luiz Koodi Hotta

University of Campinas (UNICAMP) - Department of Statistics

Pedro L. Valls Pereira

Sao Paulo School of Economics - FGV and CEQEF- FGV

Mauricio Zevallos

Universidade Estadual de Campinas (UNICAMP)

Date Written: September 26, 2020

Abstract

Based on a General Dynamic Factor Model with infinite-dimensional factor space and MGARCH common shocks, we develop new estimation and forecasting procedures for conditional covariance matrices in high-dimensional time series. The finite-sample performance of our approach is evaluated via Monte Carlo experiments, outperforming most alternative methods. The new procedure is used to construct one-step-ahead minimum variance portfolios for a high-dimensional panel of assets. The results are shown to achieve better out-of-sample portfolio performance than alternative existing procedures.

Keywords: Dimension reduction, Large panels, High-dimensional time series, Minimum variance portfolio, Volatility, Multivariate GARCH

JEL Classification: C38, C53, C55, C59, G11

Suggested Citation

Trucíos Maza, Carlos César and Mazzeu, João Henrique Gonçalves and Hallin, Marc and Hotta, Luiz Koodi and Valls Pereira, Pedro L. and Zevallos, Mauricio, Forecasting Conditional Covariance Matrices in High-Dimensional Time Series: A General Dynamic Factor Approach (September 26, 2020). Available at SSRN: https://ssrn.com/abstract=3399782 or http://dx.doi.org/10.2139/ssrn.3399782

Carlos César Trucíos Maza (Contact Author)

Universidade Federal do Rio de Janeiro (UFRJ) - Faculdade de Administracao e Ciencias Contabeis - FACC ( email )

Av. Pasteur 250
Rio de Janeiro, 22290-240
Brazil

HOME PAGE: http://https://ctruciosm.github.io

João Henrique Gonçalves Mazzeu

Universidad Carlos III de Madrid

CL. de Madrid 126
Madrid, 28903
Spain

Marc Hallin

ECARES, Universite Libre de Bruxelles ( email )

Ave. Franklin D Roosevelt, 50 - C.P. 114
Brussels, B-1050
Belgium
+32 2 650 5886 (Phone)
+32 2 650 5899 (Fax)

Luiz Koodi Hotta

University of Campinas (UNICAMP) - Department of Statistics ( email )

Campinas, São Paulo 13083-859
Brazil

Pedro L. Valls Pereira

Sao Paulo School of Economics - FGV and CEQEF- FGV ( email )

Rua Itapeva 474 room 1006
São Paulo, São Paulo 01332-000
Brazil
55+11+37993726 (Phone)
55+11+37993357 (Fax)

HOME PAGE: http://sites.google.com/site/pedrovallspereira

Mauricio Zevallos

Universidade Estadual de Campinas (UNICAMP) ( email )

Rua Sérgio Buarque de Holanda, 651
Cidade Universitaria, Barao Geraldo
Campinas, Sao Paulo
Brazil

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
42
Abstract Views
376
PlumX Metrics