Testing for equal predictive accuracy with strong dependence

Laura Coroneo; Fabrizio Iacone

Testing for equal predictive accuracy with strong dependence

Laura Coroneo, Fabrizio Iacone^*

^*Corresponding author for this work

Economics

Research output: Contribution to journal › Article › peer-review

Abstract

We analyse the properties of the Diebold and Mariano (1995) test in the presence of autocorrelation in the loss differential. We show that the power of the Diebold and Mariano (1995) test decreases as the dependence increases, making it more difficult to obtain statistically significant evidence of superior predictive ability against less accurate benchmarks. We also find that, after a certain threshold, the test has no power and the correct null hypothesis is spuriously rejected. Taken together, these results caution to seriously consider the dependence properties of the loss differential before the application of the Diebold and Mariano (1995) test.

Original language	English
Journal	International journal of forecasting
Publication status	Accepted/In press - 4 Nov 2024

Bibliographical note

This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.

Access to Document

2409.12662v1
This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.
Accepted author manuscript, 496 KBLicence: CC BY

https://arxiv.org/abs/2409.12662Licence: Unspecified

Cite this

@article{9654d4f748fc4994af60375ec6f342d3,

title = "Testing for equal predictive accuracy with strong dependence",

abstract = "We analyse the properties of the Diebold and Mariano (1995) test in the presence of autocorrelation in the loss differential. We show that the power of the Diebold and Mariano (1995) test decreases as the dependence increases, making it more difficult to obtain statistically significant evidence of superior predictive ability against less accurate benchmarks. We also find that, after a certain threshold, the test has no power and the correct null hypothesis is spuriously rejected. Taken together, these results caution to seriously consider the dependence properties of the loss differential before the application of the Diebold and Mariano (1995) test.",

author = "Laura Coroneo and Fabrizio Iacone",

note = "This is an author-produced version of the published paper. Uploaded in accordance with the University{\textquoteright}s Research Publications and Open Access policy.",

year = "2024",

month = nov,

day = "4",

language = "English",

journal = "International journal of forecasting",

issn = "0169-2070",

publisher = "Elsevier",

}

TY - JOUR

T1 - Testing for equal predictive accuracy with strong dependence

AU - Coroneo, Laura

AU - Iacone, Fabrizio

N1 - This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.

PY - 2024/11/4

Y1 - 2024/11/4

N2 - We analyse the properties of the Diebold and Mariano (1995) test in the presence of autocorrelation in the loss differential. We show that the power of the Diebold and Mariano (1995) test decreases as the dependence increases, making it more difficult to obtain statistically significant evidence of superior predictive ability against less accurate benchmarks. We also find that, after a certain threshold, the test has no power and the correct null hypothesis is spuriously rejected. Taken together, these results caution to seriously consider the dependence properties of the loss differential before the application of the Diebold and Mariano (1995) test.

AB - We analyse the properties of the Diebold and Mariano (1995) test in the presence of autocorrelation in the loss differential. We show that the power of the Diebold and Mariano (1995) test decreases as the dependence increases, making it more difficult to obtain statistically significant evidence of superior predictive ability against less accurate benchmarks. We also find that, after a certain threshold, the test has no power and the correct null hypothesis is spuriously rejected. Taken together, these results caution to seriously consider the dependence properties of the loss differential before the application of the Diebold and Mariano (1995) test.

M3 - Article

SN - 0169-2070

JO - International journal of forecasting

JF - International journal of forecasting

ER -