Gresham’s Law of model averaging

In-Koo Cho, Kenneth Kasa

Research output: Contribution to journalReview article

Abstract

A decision maker doubts the stationarity of his environment. In response, he uses two models, one with time-varying parameters, and another with constant parameters. Forecasts are then based on a Bayesian model averaging strategy, which mixes forecasts from the two models. In reality, structural parameters are constant, but the (unknown) true model features expectational feedback, which the reduced-form models neglect. This feedback permits fears of parameter instability to become self-confirming. Within the context of a standard asset-pricing model, we use the tools of large deviations theory to show that even though the constant parameter model would converge to the rational expectations equilibrium if considered in isolation, the mere presence of an unstable alternative drives it out of consideration.

Original languageEnglish (US)
Pages (from-to)3589-3616
Number of pages28
JournalAmerican Economic Review
Volume107
Issue number11
DOIs
StatePublished - Nov 2017

Fingerprint

Model averaging
Asset pricing models
Rational expectations equilibrium
Reduced-form model
Parameter instability
Bayesian model averaging
Neglect
Time-varying parameters
Isolation
Large deviations
Stationarity
Structural parameters
Decision maker

ASJC Scopus subject areas

  • Economics and Econometrics

Cite this

Gresham’s Law of model averaging. / Cho, In-Koo; Kasa, Kenneth.

In: American Economic Review, Vol. 107, No. 11, 11.2017, p. 3589-3616.

Research output: Contribution to journalReview article

Cho, I-K & Kasa, K 2017, 'Gresham’s Law of model averaging', American Economic Review, vol. 107, no. 11, pp. 3589-3616. https://doi.org/10.1257/aer.20160665
Cho, In-Koo ; Kasa, Kenneth. / Gresham’s Law of model averaging. In: American Economic Review. 2017 ; Vol. 107, No. 11. pp. 3589-3616.
@article{d933409caf8446008c53ad0d6588aef3,
title = "Gresham’s Law of model averaging",
abstract = "A decision maker doubts the stationarity of his environment. In response, he uses two models, one with time-varying parameters, and another with constant parameters. Forecasts are then based on a Bayesian model averaging strategy, which mixes forecasts from the two models. In reality, structural parameters are constant, but the (unknown) true model features expectational feedback, which the reduced-form models neglect. This feedback permits fears of parameter instability to become self-confirming. Within the context of a standard asset-pricing model, we use the tools of large deviations theory to show that even though the constant parameter model would converge to the rational expectations equilibrium if considered in isolation, the mere presence of an unstable alternative drives it out of consideration.",
author = "In-Koo Cho and Kenneth Kasa",
year = "2017",
month = "11",
doi = "10.1257/aer.20160665",
language = "English (US)",
volume = "107",
pages = "3589--3616",
journal = "American Economic Review",
issn = "0002-8282",
publisher = "American Economic Association",
number = "11",

}

TY - JOUR

T1 - Gresham’s Law of model averaging

AU - Cho, In-Koo

AU - Kasa, Kenneth

PY - 2017/11

Y1 - 2017/11

N2 - A decision maker doubts the stationarity of his environment. In response, he uses two models, one with time-varying parameters, and another with constant parameters. Forecasts are then based on a Bayesian model averaging strategy, which mixes forecasts from the two models. In reality, structural parameters are constant, but the (unknown) true model features expectational feedback, which the reduced-form models neglect. This feedback permits fears of parameter instability to become self-confirming. Within the context of a standard asset-pricing model, we use the tools of large deviations theory to show that even though the constant parameter model would converge to the rational expectations equilibrium if considered in isolation, the mere presence of an unstable alternative drives it out of consideration.

AB - A decision maker doubts the stationarity of his environment. In response, he uses two models, one with time-varying parameters, and another with constant parameters. Forecasts are then based on a Bayesian model averaging strategy, which mixes forecasts from the two models. In reality, structural parameters are constant, but the (unknown) true model features expectational feedback, which the reduced-form models neglect. This feedback permits fears of parameter instability to become self-confirming. Within the context of a standard asset-pricing model, we use the tools of large deviations theory to show that even though the constant parameter model would converge to the rational expectations equilibrium if considered in isolation, the mere presence of an unstable alternative drives it out of consideration.

UR - http://www.scopus.com/inward/record.url?scp=85032227693&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85032227693&partnerID=8YFLogxK

U2 - 10.1257/aer.20160665

DO - 10.1257/aer.20160665

M3 - Review article

AN - SCOPUS:85032227693

VL - 107

SP - 3589

EP - 3616

JO - American Economic Review

JF - American Economic Review

SN - 0002-8282

IS - 11

ER -