Forecast density combinations of dynamic models and data driven portfolio strategies

N. Baştürk, A. Borowska, S. Grassi, L. Hoogerheide, H. K. van Dijk

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

A dynamic asset-allocation model is specified in probabilistic terms as a combination of return distributions resulting from multiple pairs of dynamic models and portfolio strategies based on momentum patterns in US industry returns. The nonlinear state space representation of the model allows efficient and robust simulation-based Bayesian inference using a novel non-linear filter. Combination weights can be cross-correlated and correlated over time using feedback mechanisms. Diagnostic analysis gives insight into model and strategy misspecification. Empirical results show that a smaller flexible model-strategy combination performs better in terms of expected return and risk than a larger basic model-strategy combination. Dynamic patterns in combination weights and diagnostic learning provide useful signals for improved modeling and policy, in particular, from a risk-management perspective.

Original languageEnglish
Pages (from-to)170-186
JournalJournal of Econometrics
Volume210
Issue number1
DOIs
Publication statusPublished - May 2019

Fingerprint

Portfolio strategy
Density forecasts
Diagnostics
Empirical results
Modeling
Dynamic asset allocation
Bayesian inference
Filter
State space
Expected returns
Misspecification
Industry
Return distribution
Feedback mechanism
Risk management
Momentum
Simulation

Keywords

  • Bayes estimates
  • Filtering methods
  • Forecast combination
  • Momentum strategy

Cite this

Baştürk, N. ; Borowska, A. ; Grassi, S. ; Hoogerheide, L. ; van Dijk, H. K. / Forecast density combinations of dynamic models and data driven portfolio strategies. In: Journal of Econometrics. 2019 ; Vol. 210, No. 1. pp. 170-186.
@article{06914669305c4570809b495ab24e300a,
title = "Forecast density combinations of dynamic models and data driven portfolio strategies",
abstract = "A dynamic asset-allocation model is specified in probabilistic terms as a combination of return distributions resulting from multiple pairs of dynamic models and portfolio strategies based on momentum patterns in US industry returns. The nonlinear state space representation of the model allows efficient and robust simulation-based Bayesian inference using a novel non-linear filter. Combination weights can be cross-correlated and correlated over time using feedback mechanisms. Diagnostic analysis gives insight into model and strategy misspecification. Empirical results show that a smaller flexible model-strategy combination performs better in terms of expected return and risk than a larger basic model-strategy combination. Dynamic patterns in combination weights and diagnostic learning provide useful signals for improved modeling and policy, in particular, from a risk-management perspective.",
keywords = "Bayes estimates, Filtering methods, Forecast combination, Momentum strategy",
author = "N. Başt{\"u}rk and A. Borowska and S. Grassi and L. Hoogerheide and {van Dijk}, {H. K.}",
year = "2019",
month = "5",
doi = "10.1016/j.jeconom.2018.11.011",
language = "English",
volume = "210",
pages = "170--186",
journal = "Journal of Econometrics",
issn = "0304-4076",
publisher = "Elsevier BV",
number = "1",

}

Forecast density combinations of dynamic models and data driven portfolio strategies. / Baştürk, N.; Borowska, A.; Grassi, S.; Hoogerheide, L.; van Dijk, H. K.

In: Journal of Econometrics, Vol. 210, No. 1, 05.2019, p. 170-186.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Forecast density combinations of dynamic models and data driven portfolio strategies

AU - Baştürk, N.

AU - Borowska, A.

AU - Grassi, S.

AU - Hoogerheide, L.

AU - van Dijk, H. K.

PY - 2019/5

Y1 - 2019/5

N2 - A dynamic asset-allocation model is specified in probabilistic terms as a combination of return distributions resulting from multiple pairs of dynamic models and portfolio strategies based on momentum patterns in US industry returns. The nonlinear state space representation of the model allows efficient and robust simulation-based Bayesian inference using a novel non-linear filter. Combination weights can be cross-correlated and correlated over time using feedback mechanisms. Diagnostic analysis gives insight into model and strategy misspecification. Empirical results show that a smaller flexible model-strategy combination performs better in terms of expected return and risk than a larger basic model-strategy combination. Dynamic patterns in combination weights and diagnostic learning provide useful signals for improved modeling and policy, in particular, from a risk-management perspective.

AB - A dynamic asset-allocation model is specified in probabilistic terms as a combination of return distributions resulting from multiple pairs of dynamic models and portfolio strategies based on momentum patterns in US industry returns. The nonlinear state space representation of the model allows efficient and robust simulation-based Bayesian inference using a novel non-linear filter. Combination weights can be cross-correlated and correlated over time using feedback mechanisms. Diagnostic analysis gives insight into model and strategy misspecification. Empirical results show that a smaller flexible model-strategy combination performs better in terms of expected return and risk than a larger basic model-strategy combination. Dynamic patterns in combination weights and diagnostic learning provide useful signals for improved modeling and policy, in particular, from a risk-management perspective.

KW - Bayes estimates

KW - Filtering methods

KW - Forecast combination

KW - Momentum strategy

UR - http://www.scopus.com/inward/record.url?scp=85057214511&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057214511&partnerID=8YFLogxK

U2 - 10.1016/j.jeconom.2018.11.011

DO - 10.1016/j.jeconom.2018.11.011

M3 - Article

VL - 210

SP - 170

EP - 186

JO - Journal of Econometrics

JF - Journal of Econometrics

SN - 0304-4076

IS - 1

ER -