Forecasting With Model Uncertainty: Representations and Risk Reduction

Keisuke Hirano, Jonathan H. Wright

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

We consider forecasting with uncertainty about the choice of predictor variables. The researcher wants to select a model, estimate the parameters, and use the parameter estimates for forecasting. We investigate the distributional properties of a number of different schemes for model choice and parameter estimation, including: in-sample model selection using the Akaike information criterion; out-of-sample model selection; and splitting the data into subsamples for model selection and parameter estimation. Using a weak-predictor local asymptotic scheme, we provide a representation result that facilitates comparison of the distributional properties of the procedures and their associated forecast risks. This representation isolates the source of inefficiency in some of these procedures. We develop a simulation procedure that improves the accuracy of the out-of-sample and split-sample methods uniformly over the local parameter space. We also examine how bootstrap aggregation (bagging) affects the local asymptotic risk of the estimators and their associated forecasts. Numerically, we find that for many values of the local parameter, the out-of-sample and split-sample schemes perform poorly if implemented in the conventional way. But they perform well, if implemented in conjunction with our risk-reduction method or bagging.

Original languageEnglish (US)
Pages (from-to)617-643
Number of pages27
JournalEconometrica
Volume85
Issue number2
DOIs
StatePublished - Mar 1 2017

Fingerprint

Model uncertainty
Risk reduction
Sample selection model
Parameter estimation
Bagging
Predictors
Model choice
Akaike information criterion
Simulation
Uncertainty
Model selection
Inefficiency
Bootstrap
Estimator

All Science Journal Classification (ASJC) codes

  • Economics and Econometrics

Cite this

Hirano, Keisuke ; Wright, Jonathan H. / Forecasting With Model Uncertainty : Representations and Risk Reduction. In: Econometrica. 2017 ; Vol. 85, No. 2. pp. 617-643.
@article{aa62579a50e241209b18651291cf1eba,
title = "Forecasting With Model Uncertainty: Representations and Risk Reduction",
abstract = "We consider forecasting with uncertainty about the choice of predictor variables. The researcher wants to select a model, estimate the parameters, and use the parameter estimates for forecasting. We investigate the distributional properties of a number of different schemes for model choice and parameter estimation, including: in-sample model selection using the Akaike information criterion; out-of-sample model selection; and splitting the data into subsamples for model selection and parameter estimation. Using a weak-predictor local asymptotic scheme, we provide a representation result that facilitates comparison of the distributional properties of the procedures and their associated forecast risks. This representation isolates the source of inefficiency in some of these procedures. We develop a simulation procedure that improves the accuracy of the out-of-sample and split-sample methods uniformly over the local parameter space. We also examine how bootstrap aggregation (bagging) affects the local asymptotic risk of the estimators and their associated forecasts. Numerically, we find that for many values of the local parameter, the out-of-sample and split-sample schemes perform poorly if implemented in the conventional way. But they perform well, if implemented in conjunction with our risk-reduction method or bagging.",
author = "Keisuke Hirano and Wright, {Jonathan H.}",
year = "2017",
month = "3",
day = "1",
doi = "10.3982/ECTA13372",
language = "English (US)",
volume = "85",
pages = "617--643",
journal = "Econometrica",
issn = "0012-9682",
publisher = "Wiley-Blackwell",
number = "2",

}

Forecasting With Model Uncertainty : Representations and Risk Reduction. / Hirano, Keisuke; Wright, Jonathan H.

In: Econometrica, Vol. 85, No. 2, 01.03.2017, p. 617-643.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Forecasting With Model Uncertainty

T2 - Representations and Risk Reduction

AU - Hirano, Keisuke

AU - Wright, Jonathan H.

PY - 2017/3/1

Y1 - 2017/3/1

N2 - We consider forecasting with uncertainty about the choice of predictor variables. The researcher wants to select a model, estimate the parameters, and use the parameter estimates for forecasting. We investigate the distributional properties of a number of different schemes for model choice and parameter estimation, including: in-sample model selection using the Akaike information criterion; out-of-sample model selection; and splitting the data into subsamples for model selection and parameter estimation. Using a weak-predictor local asymptotic scheme, we provide a representation result that facilitates comparison of the distributional properties of the procedures and their associated forecast risks. This representation isolates the source of inefficiency in some of these procedures. We develop a simulation procedure that improves the accuracy of the out-of-sample and split-sample methods uniformly over the local parameter space. We also examine how bootstrap aggregation (bagging) affects the local asymptotic risk of the estimators and their associated forecasts. Numerically, we find that for many values of the local parameter, the out-of-sample and split-sample schemes perform poorly if implemented in the conventional way. But they perform well, if implemented in conjunction with our risk-reduction method or bagging.

AB - We consider forecasting with uncertainty about the choice of predictor variables. The researcher wants to select a model, estimate the parameters, and use the parameter estimates for forecasting. We investigate the distributional properties of a number of different schemes for model choice and parameter estimation, including: in-sample model selection using the Akaike information criterion; out-of-sample model selection; and splitting the data into subsamples for model selection and parameter estimation. Using a weak-predictor local asymptotic scheme, we provide a representation result that facilitates comparison of the distributional properties of the procedures and their associated forecast risks. This representation isolates the source of inefficiency in some of these procedures. We develop a simulation procedure that improves the accuracy of the out-of-sample and split-sample methods uniformly over the local parameter space. We also examine how bootstrap aggregation (bagging) affects the local asymptotic risk of the estimators and their associated forecasts. Numerically, we find that for many values of the local parameter, the out-of-sample and split-sample schemes perform poorly if implemented in the conventional way. But they perform well, if implemented in conjunction with our risk-reduction method or bagging.

UR - http://www.scopus.com/inward/record.url?scp=85015886199&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015886199&partnerID=8YFLogxK

U2 - 10.3982/ECTA13372

DO - 10.3982/ECTA13372

M3 - Article

AN - SCOPUS:85015886199

VL - 85

SP - 617

EP - 643

JO - Econometrica

JF - Econometrica

SN - 0012-9682

IS - 2

ER -