A new semiparametric test for superior predictive ability

Zongwu Cai, Jiancheng Jiang, Jingshuang Zhang, Xibin Zhang

Research output: Contribution to journalArticleResearchpeer-review

Abstract

We propose a new method to test the superior predictive ability (SPA) of a benchmark model against a large group of alternative models. The proposed test is useful for reducing potential data snooping bias. Unlike previous methods, we model the covariance matrix by factor models and develop a generalized likelihood ratio (GLR) test statistic for the above testing problem. The GLR test is also extended to a stepwise GLR (step-GLR) test in the spirit of the step-RC test of Romano and Wolf (Econometrica 73(4):1237-1282, 2005) and step-SPA test of Hsu et al. (J Empir Financ 17(3):471-484, 2010). The step-GLR test can identify the most contributed predictive models to the rejection of the null hypothesis. A Monte Carlo simulation study shows that the GLR test is much more powerful and less conservative than the SPA test of Hansen (J Bus Econ Stat 23(4):365-380, 2005). We also present an application to illustrate the use of the GLR test and make a comparison between our GLR and Hansen s SPA tests.
Original languageEnglish
Pages (from-to)389 - 405
Number of pages17
JournalEmpirical Economics
Volume48
Issue number1
DOIs
Publication statusPublished - 2015

Cite this

Cai, Zongwu ; Jiang, Jiancheng ; Zhang, Jingshuang ; Zhang, Xibin. / A new semiparametric test for superior predictive ability. In: Empirical Economics. 2015 ; Vol. 48, No. 1. pp. 389 - 405.
@article{998e6d002c184f3f8d911bbe9fab8bd3,
title = "A new semiparametric test for superior predictive ability",
abstract = "We propose a new method to test the superior predictive ability (SPA) of a benchmark model against a large group of alternative models. The proposed test is useful for reducing potential data snooping bias. Unlike previous methods, we model the covariance matrix by factor models and develop a generalized likelihood ratio (GLR) test statistic for the above testing problem. The GLR test is also extended to a stepwise GLR (step-GLR) test in the spirit of the step-RC test of Romano and Wolf (Econometrica 73(4):1237-1282, 2005) and step-SPA test of Hsu et al. (J Empir Financ 17(3):471-484, 2010). The step-GLR test can identify the most contributed predictive models to the rejection of the null hypothesis. A Monte Carlo simulation study shows that the GLR test is much more powerful and less conservative than the SPA test of Hansen (J Bus Econ Stat 23(4):365-380, 2005). We also present an application to illustrate the use of the GLR test and make a comparison between our GLR and Hansen s SPA tests.",
author = "Zongwu Cai and Jiancheng Jiang and Jingshuang Zhang and Xibin Zhang",
year = "2015",
doi = "10.1007/s00181-014-0887-6",
language = "English",
volume = "48",
pages = "389 -- 405",
journal = "Empirical Economics",
issn = "0377-7332",
publisher = "Springer-Verlag London Ltd.",
number = "1",

}

A new semiparametric test for superior predictive ability. / Cai, Zongwu; Jiang, Jiancheng; Zhang, Jingshuang; Zhang, Xibin.

In: Empirical Economics, Vol. 48, No. 1, 2015, p. 389 - 405.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - A new semiparametric test for superior predictive ability

AU - Cai, Zongwu

AU - Jiang, Jiancheng

AU - Zhang, Jingshuang

AU - Zhang, Xibin

PY - 2015

Y1 - 2015

N2 - We propose a new method to test the superior predictive ability (SPA) of a benchmark model against a large group of alternative models. The proposed test is useful for reducing potential data snooping bias. Unlike previous methods, we model the covariance matrix by factor models and develop a generalized likelihood ratio (GLR) test statistic for the above testing problem. The GLR test is also extended to a stepwise GLR (step-GLR) test in the spirit of the step-RC test of Romano and Wolf (Econometrica 73(4):1237-1282, 2005) and step-SPA test of Hsu et al. (J Empir Financ 17(3):471-484, 2010). The step-GLR test can identify the most contributed predictive models to the rejection of the null hypothesis. A Monte Carlo simulation study shows that the GLR test is much more powerful and less conservative than the SPA test of Hansen (J Bus Econ Stat 23(4):365-380, 2005). We also present an application to illustrate the use of the GLR test and make a comparison between our GLR and Hansen s SPA tests.

AB - We propose a new method to test the superior predictive ability (SPA) of a benchmark model against a large group of alternative models. The proposed test is useful for reducing potential data snooping bias. Unlike previous methods, we model the covariance matrix by factor models and develop a generalized likelihood ratio (GLR) test statistic for the above testing problem. The GLR test is also extended to a stepwise GLR (step-GLR) test in the spirit of the step-RC test of Romano and Wolf (Econometrica 73(4):1237-1282, 2005) and step-SPA test of Hsu et al. (J Empir Financ 17(3):471-484, 2010). The step-GLR test can identify the most contributed predictive models to the rejection of the null hypothesis. A Monte Carlo simulation study shows that the GLR test is much more powerful and less conservative than the SPA test of Hansen (J Bus Econ Stat 23(4):365-380, 2005). We also present an application to illustrate the use of the GLR test and make a comparison between our GLR and Hansen s SPA tests.

U2 - 10.1007/s00181-014-0887-6

DO - 10.1007/s00181-014-0887-6

M3 - Article

VL - 48

SP - 389

EP - 405

JO - Empirical Economics

JF - Empirical Economics

SN - 0377-7332

IS - 1

ER -