Inducing sparsity and shrinkage in time-varying parameter models

Florian Huber, Gary Koop, Luca Onorante

Research output: Contribution to journalArticle

Abstract

Time-varying parameter (TVP) models have the potential to be over-parameterized, particularly when the number of variables in the model is large. Global-local priors are increasingly used to induce shrinkage in such models. But the estimates produced by these priors can still have appreciable uncertainty. Sparsification has the potential to reduce this uncertainty and improve forecasts. In this paper, we develop computationally simple methods which both shrink and sparsify TVP models. In a simulated data exercise we show the benefits of our shrink-then-sparsify approach in a variety of sparse and dense TVP regressions. In a macroeconomic forecasting exercise, we find our approach to substantially improve forecast performance relative to shrinkage alone.
Original languageEnglish
Number of pages34
JournalJournal of Business and Economic Statistics
Publication statusAccepted/In press - 16 Dec 2019

Fingerprint

Time-varying Parameters
Shrinkage
Sparsity
Exercise
Forecast
uncertainty
Uncertainty
Macroeconomics
Model
macroeconomics
Forecasting
Regression
regression
time
Time-varying parameter model
Estimate
performance
Time-varying parameters
Macroeconomic forecasting
Forecast performance

Keywords

  • sparsity
  • shrinkage
  • hierarchical priors
  • time varying parameter regression

Cite this

@article{b0f54a7f63b14103924ac891406fb63b,
title = "Inducing sparsity and shrinkage in time-varying parameter models",
abstract = "Time-varying parameter (TVP) models have the potential to be over-parameterized, particularly when the number of variables in the model is large. Global-local priors are increasingly used to induce shrinkage in such models. But the estimates produced by these priors can still have appreciable uncertainty. Sparsification has the potential to reduce this uncertainty and improve forecasts. In this paper, we develop computationally simple methods which both shrink and sparsify TVP models. In a simulated data exercise we show the benefits of our shrink-then-sparsify approach in a variety of sparse and dense TVP regressions. In a macroeconomic forecasting exercise, we find our approach to substantially improve forecast performance relative to shrinkage alone.",
keywords = "sparsity, shrinkage, hierarchical priors, time varying parameter regression",
author = "Florian Huber and Gary Koop and Luca Onorante",
year = "2019",
month = "12",
day = "16",
language = "English",
journal = "Journal of Business and Economic Statistics",
issn = "0735-0015",

}

Inducing sparsity and shrinkage in time-varying parameter models. / Huber, Florian; Koop, Gary; Onorante, Luca.

In: Journal of Business and Economic Statistics, 16.12.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Inducing sparsity and shrinkage in time-varying parameter models

AU - Huber, Florian

AU - Koop, Gary

AU - Onorante, Luca

PY - 2019/12/16

Y1 - 2019/12/16

N2 - Time-varying parameter (TVP) models have the potential to be over-parameterized, particularly when the number of variables in the model is large. Global-local priors are increasingly used to induce shrinkage in such models. But the estimates produced by these priors can still have appreciable uncertainty. Sparsification has the potential to reduce this uncertainty and improve forecasts. In this paper, we develop computationally simple methods which both shrink and sparsify TVP models. In a simulated data exercise we show the benefits of our shrink-then-sparsify approach in a variety of sparse and dense TVP regressions. In a macroeconomic forecasting exercise, we find our approach to substantially improve forecast performance relative to shrinkage alone.

AB - Time-varying parameter (TVP) models have the potential to be over-parameterized, particularly when the number of variables in the model is large. Global-local priors are increasingly used to induce shrinkage in such models. But the estimates produced by these priors can still have appreciable uncertainty. Sparsification has the potential to reduce this uncertainty and improve forecasts. In this paper, we develop computationally simple methods which both shrink and sparsify TVP models. In a simulated data exercise we show the benefits of our shrink-then-sparsify approach in a variety of sparse and dense TVP regressions. In a macroeconomic forecasting exercise, we find our approach to substantially improve forecast performance relative to shrinkage alone.

KW - sparsity

KW - shrinkage

KW - hierarchical priors

KW - time varying parameter regression

UR - https://amstat.tandfonline.com/toc/ubes20/current

M3 - Article

JO - Journal of Business and Economic Statistics

JF - Journal of Business and Economic Statistics

SN - 0735-0015

ER -