Empirical bayesian density forecasting in iowa and shrinkage for the monte carlo era

Kurt F. Lewis, Charles H. Whiteman

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

The track record of a 20-year history of density forecasts of state tax revenue in Iowa is studied, and potential improvements sought through a search for better-performing 'priors' similar to that conducted three decades ago for point forecasts by Doan, Litterman and Sims (Econometric Reviews, 1984). Comparisons of the point and density forecasts produced under the flat prior are made to those produced by the traditional (mixed estimation) 'Bayesian VAR' methods of Doan, Litterman and Sims, as well as to fully Bayesian 'Minnesota Prior' forecasts. The actual record and, to a somewhat lesser extent, the record of the alternative procedures studied in pseudo-real-time forecasting experiments, share a characteristic: subsequently realized revenues are in the lower tails of the predicted distributions 'too often'. An alternative empirically based prior is found by working directly on the probability distribution for the vector autoregression parameters - the goal being to discover a better-performing entropically tilted prior that minimizes out-of-sample mean squared error subject to a Kullback-Leibler divergence constraint that the new prior not differ 'too much' from the original. We also study the closely related topic of robust prediction appropriate for situations of ambiguity. Robust 'priors' are competitive in out-of-sample forecasting; despite the freedom afforded the entropically tilted prior, it does not perform better than the simple alternatives.

Original languageEnglish (US)
Pages (from-to)15-35
Number of pages21
JournalJournal of Forecasting
Volume34
Issue number1
DOIs
StatePublished - Jan 1 2015

Fingerprint

Shrinkage
Forecast
Forecasting
Alternatives
Taxation
Vector Autoregression
Probability distributions
Kullback-Leibler Divergence
Sample mean
Tax
Bayesian Methods
Econometrics
History
Mean Squared Error
Tail
Probability Distribution
Real-time
Minimise
Density forecasting
Prediction

All Science Journal Classification (ASJC) codes

  • Modeling and Simulation
  • Computer Science Applications
  • Strategy and Management
  • Statistics, Probability and Uncertainty
  • Management Science and Operations Research

Cite this

@article{4c283a42b9d8497a9974ceb3a64652e2,
title = "Empirical bayesian density forecasting in iowa and shrinkage for the monte carlo era",
abstract = "The track record of a 20-year history of density forecasts of state tax revenue in Iowa is studied, and potential improvements sought through a search for better-performing 'priors' similar to that conducted three decades ago for point forecasts by Doan, Litterman and Sims (Econometric Reviews, 1984). Comparisons of the point and density forecasts produced under the flat prior are made to those produced by the traditional (mixed estimation) 'Bayesian VAR' methods of Doan, Litterman and Sims, as well as to fully Bayesian 'Minnesota Prior' forecasts. The actual record and, to a somewhat lesser extent, the record of the alternative procedures studied in pseudo-real-time forecasting experiments, share a characteristic: subsequently realized revenues are in the lower tails of the predicted distributions 'too often'. An alternative empirically based prior is found by working directly on the probability distribution for the vector autoregression parameters - the goal being to discover a better-performing entropically tilted prior that minimizes out-of-sample mean squared error subject to a Kullback-Leibler divergence constraint that the new prior not differ 'too much' from the original. We also study the closely related topic of robust prediction appropriate for situations of ambiguity. Robust 'priors' are competitive in out-of-sample forecasting; despite the freedom afforded the entropically tilted prior, it does not perform better than the simple alternatives.",
author = "Lewis, {Kurt F.} and Whiteman, {Charles H.}",
year = "2015",
month = "1",
day = "1",
doi = "10.1002/for.2316",
language = "English (US)",
volume = "34",
pages = "15--35",
journal = "Journal of Forecasting",
issn = "0277-6693",
publisher = "John Wiley and Sons Ltd",
number = "1",

}

Empirical bayesian density forecasting in iowa and shrinkage for the monte carlo era. / Lewis, Kurt F.; Whiteman, Charles H.

In: Journal of Forecasting, Vol. 34, No. 1, 01.01.2015, p. 15-35.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Empirical bayesian density forecasting in iowa and shrinkage for the monte carlo era

AU - Lewis, Kurt F.

AU - Whiteman, Charles H.

PY - 2015/1/1

Y1 - 2015/1/1

N2 - The track record of a 20-year history of density forecasts of state tax revenue in Iowa is studied, and potential improvements sought through a search for better-performing 'priors' similar to that conducted three decades ago for point forecasts by Doan, Litterman and Sims (Econometric Reviews, 1984). Comparisons of the point and density forecasts produced under the flat prior are made to those produced by the traditional (mixed estimation) 'Bayesian VAR' methods of Doan, Litterman and Sims, as well as to fully Bayesian 'Minnesota Prior' forecasts. The actual record and, to a somewhat lesser extent, the record of the alternative procedures studied in pseudo-real-time forecasting experiments, share a characteristic: subsequently realized revenues are in the lower tails of the predicted distributions 'too often'. An alternative empirically based prior is found by working directly on the probability distribution for the vector autoregression parameters - the goal being to discover a better-performing entropically tilted prior that minimizes out-of-sample mean squared error subject to a Kullback-Leibler divergence constraint that the new prior not differ 'too much' from the original. We also study the closely related topic of robust prediction appropriate for situations of ambiguity. Robust 'priors' are competitive in out-of-sample forecasting; despite the freedom afforded the entropically tilted prior, it does not perform better than the simple alternatives.

AB - The track record of a 20-year history of density forecasts of state tax revenue in Iowa is studied, and potential improvements sought through a search for better-performing 'priors' similar to that conducted three decades ago for point forecasts by Doan, Litterman and Sims (Econometric Reviews, 1984). Comparisons of the point and density forecasts produced under the flat prior are made to those produced by the traditional (mixed estimation) 'Bayesian VAR' methods of Doan, Litterman and Sims, as well as to fully Bayesian 'Minnesota Prior' forecasts. The actual record and, to a somewhat lesser extent, the record of the alternative procedures studied in pseudo-real-time forecasting experiments, share a characteristic: subsequently realized revenues are in the lower tails of the predicted distributions 'too often'. An alternative empirically based prior is found by working directly on the probability distribution for the vector autoregression parameters - the goal being to discover a better-performing entropically tilted prior that minimizes out-of-sample mean squared error subject to a Kullback-Leibler divergence constraint that the new prior not differ 'too much' from the original. We also study the closely related topic of robust prediction appropriate for situations of ambiguity. Robust 'priors' are competitive in out-of-sample forecasting; despite the freedom afforded the entropically tilted prior, it does not perform better than the simple alternatives.

UR - http://www.scopus.com/inward/record.url?scp=84919952722&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84919952722&partnerID=8YFLogxK

U2 - 10.1002/for.2316

DO - 10.1002/for.2316

M3 - Article

VL - 34

SP - 15

EP - 35

JO - Journal of Forecasting

JF - Journal of Forecasting

SN - 0277-6693

IS - 1

ER -