Best predictive small area estimation

Jiming Jiang, Thuan Nguyen, Jonnagadda S Rao

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

We derive the best predictive estimator (BPE) of the fixed parameters under two well-known small area models, the Fay-Herriot model and the nested-error regression model. This leads to a new prediction procedure, called observed best prediction (OBP), which is different from the empirical best linear unbiased prediction (EBLUP). We show that BPE is more reasonable than the traditional estimators derived from estimation considerations, such as maximum likelihood (ML) and restricted maximum likelihood (REML), if the main interest is estimation of small area means, which is a mixed-model prediction problem. We use both theoretical derivations and empirical studies to demonstrate that the OBP can significantly outperform EBLUP in terms of the mean squared prediction error (MSPE), if the underlying model is misspecified. On the other hand, when the underlying model is correctly specified, the overall predictive performance of the OBP is very similar to that of the EBLUP if the number of small areas is large. A general theory about OBP, including its exact MSPE comparison with the BLUP in the context of mixed-model prediction, and asymptotic behavior of the BPE, is developed. A real data example is considered. A supplementary appendix is available online.

Original languageEnglish
Pages (from-to)732-745
Number of pages14
JournalJournal of the American Statistical Association
Volume106
Issue number494
DOIs
StatePublished - Jun 1 2011

Fingerprint

Small Area Estimation
Best Linear Unbiased Prediction
Prediction
Estimator
Mixed Model
Prediction Error
Mean Squared Error
Restricted Maximum Likelihood
Error Model
Small area estimation
Model
Empirical Study
Maximum Likelihood
Regression Model
Asymptotic Behavior

Keywords

  • Fay-herriot model
  • Mean squared prediction error (MSPE)
  • Model misspecification
  • Nested-error regression model
  • Robustness

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Best predictive small area estimation. / Jiang, Jiming; Nguyen, Thuan; Rao, Jonnagadda S.

In: Journal of the American Statistical Association, Vol. 106, No. 494, 01.06.2011, p. 732-745.

Research output: Contribution to journalArticle

Jiang, Jiming ; Nguyen, Thuan ; Rao, Jonnagadda S. / Best predictive small area estimation. In: Journal of the American Statistical Association. 2011 ; Vol. 106, No. 494. pp. 732-745.
@article{876720287adb40e9894a82a85c1f0949,
title = "Best predictive small area estimation",
abstract = "We derive the best predictive estimator (BPE) of the fixed parameters under two well-known small area models, the Fay-Herriot model and the nested-error regression model. This leads to a new prediction procedure, called observed best prediction (OBP), which is different from the empirical best linear unbiased prediction (EBLUP). We show that BPE is more reasonable than the traditional estimators derived from estimation considerations, such as maximum likelihood (ML) and restricted maximum likelihood (REML), if the main interest is estimation of small area means, which is a mixed-model prediction problem. We use both theoretical derivations and empirical studies to demonstrate that the OBP can significantly outperform EBLUP in terms of the mean squared prediction error (MSPE), if the underlying model is misspecified. On the other hand, when the underlying model is correctly specified, the overall predictive performance of the OBP is very similar to that of the EBLUP if the number of small areas is large. A general theory about OBP, including its exact MSPE comparison with the BLUP in the context of mixed-model prediction, and asymptotic behavior of the BPE, is developed. A real data example is considered. A supplementary appendix is available online.",
keywords = "Fay-herriot model, Mean squared prediction error (MSPE), Model misspecification, Nested-error regression model, Robustness",
author = "Jiming Jiang and Thuan Nguyen and Rao, {Jonnagadda S}",
year = "2011",
month = "6",
day = "1",
doi = "10.1198/jasa.2011.tm10221",
language = "English",
volume = "106",
pages = "732--745",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "494",

}

TY - JOUR

T1 - Best predictive small area estimation

AU - Jiang, Jiming

AU - Nguyen, Thuan

AU - Rao, Jonnagadda S

PY - 2011/6/1

Y1 - 2011/6/1

N2 - We derive the best predictive estimator (BPE) of the fixed parameters under two well-known small area models, the Fay-Herriot model and the nested-error regression model. This leads to a new prediction procedure, called observed best prediction (OBP), which is different from the empirical best linear unbiased prediction (EBLUP). We show that BPE is more reasonable than the traditional estimators derived from estimation considerations, such as maximum likelihood (ML) and restricted maximum likelihood (REML), if the main interest is estimation of small area means, which is a mixed-model prediction problem. We use both theoretical derivations and empirical studies to demonstrate that the OBP can significantly outperform EBLUP in terms of the mean squared prediction error (MSPE), if the underlying model is misspecified. On the other hand, when the underlying model is correctly specified, the overall predictive performance of the OBP is very similar to that of the EBLUP if the number of small areas is large. A general theory about OBP, including its exact MSPE comparison with the BLUP in the context of mixed-model prediction, and asymptotic behavior of the BPE, is developed. A real data example is considered. A supplementary appendix is available online.

AB - We derive the best predictive estimator (BPE) of the fixed parameters under two well-known small area models, the Fay-Herriot model and the nested-error regression model. This leads to a new prediction procedure, called observed best prediction (OBP), which is different from the empirical best linear unbiased prediction (EBLUP). We show that BPE is more reasonable than the traditional estimators derived from estimation considerations, such as maximum likelihood (ML) and restricted maximum likelihood (REML), if the main interest is estimation of small area means, which is a mixed-model prediction problem. We use both theoretical derivations and empirical studies to demonstrate that the OBP can significantly outperform EBLUP in terms of the mean squared prediction error (MSPE), if the underlying model is misspecified. On the other hand, when the underlying model is correctly specified, the overall predictive performance of the OBP is very similar to that of the EBLUP if the number of small areas is large. A general theory about OBP, including its exact MSPE comparison with the BLUP in the context of mixed-model prediction, and asymptotic behavior of the BPE, is developed. A real data example is considered. A supplementary appendix is available online.

KW - Fay-herriot model

KW - Mean squared prediction error (MSPE)

KW - Model misspecification

KW - Nested-error regression model

KW - Robustness

UR - http://www.scopus.com/inward/record.url?scp=79960147873&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79960147873&partnerID=8YFLogxK

U2 - 10.1198/jasa.2011.tm10221

DO - 10.1198/jasa.2011.tm10221

M3 - Article

AN - SCOPUS:79960147873

VL - 106

SP - 732

EP - 745

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 494

ER -