Conditional Superior Predictive Ability

Jia Li; Zhipeng Liao; Rogier Quaedvlieg; Jia Li

doi:10.1093/restud/rdab039

Conditional Superior Predictive Ability

Jia Li
, Zhipeng Liao
, Rogier Quaedvlieg
, Jia Li

University of California at Los Angeles

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

26 Citations (Scopus)

549 Downloads (Pure)

Abstract

This article proposes a test for the conditional superior predictive ability (CSPA) of a family of forecasting methods with respect to a benchmark. The test is functional in nature: under the null hypothesis, the benchmark's conditional expected loss is no more than those of the competitors, uniformly across all conditioning states. By inverting the CSPA tests for a set of benchmarks, we obtain confidence sets for the uniformly most superior method. The econometric inference pertains to testing conditional moment inequalities for time series data with general serial dependence, and we justify its asymptotic validity using a uniform non-parametric inference method based on a new strong approximation theory for mixingales. The usefulness of the method is demonstrated in empirical applications on volatility and inflation forecasting.

Original language	English
Pages (from-to)	843-875
Number of pages	33
Journal	Review of Economic Studies
Volume	89
Issue number	2
DOIs	https://doi.org/10.1093/restud/rdab039
Publication status	Published - 6 Mar 2022

Bibliographical note

Acknowledgments: We thank the Co-Editor (Francesca Molinari) and four anonymous referees for their comments
and suggestions, which have greatly improved the paper. We also thank Raffaella Giacomini, Jinyong Hahn, Peter
Reinhard Hansen (discussant), Hyungsik Roger Moon and conference and seminar participants at Aarhus, CREST, the
2018 Triangle Econometrics Conference, Southern California Winter Econometrics Day, the 2019 Toulouse Financial
Econometrics Conference, and the 2021 SoFiE seminar for their comments. Liao’s research was partially supported by
National Science Foundation Grant SES-1628889. Quaedvlieg was financially supported by the Netherlands Organisation
for Scientific Research (NWO) Grant 451-17-009.

Publisher Copyright: © 2021 The Author(s) 2021. Published by Oxford University Press on behalf of The Review of Economic Studies Limited.

Funding Information:
Permission by Surface Technologies Ltd. to publish the results of this work and the help of Mr. Y. Srur and Mr. A. Ronen in texturing the seal rings and running the experiments is gratefully acknowledged. The work was partially supported by the Israel Commerce and Industry Ministry and by the Fund for the Promotion of Research at the Technion.

Publisher Copyright:
© 2021 The Author(s) 2021. Published by Oxford University Press on behalf of The Review of Economic Studies Limited.

Access to Document

10.1093/restud/rdab039

rdab039Licence: Article 25fa Dutch Copyright Act

Cite this

@article{a86320e031b54a728741bbde1bd730e2,

title = "Conditional Superior Predictive Ability",

abstract = "This article proposes a test for the conditional superior predictive ability (CSPA) of a family of forecasting methods with respect to a benchmark. The test is functional in nature: under the null hypothesis, the benchmark's conditional expected loss is no more than those of the competitors, uniformly across all conditioning states. By inverting the CSPA tests for a set of benchmarks, we obtain confidence sets for the uniformly most superior method. The econometric inference pertains to testing conditional moment inequalities for time series data with general serial dependence, and we justify its asymptotic validity using a uniform non-parametric inference method based on a new strong approximation theory for mixingales. The usefulness of the method is demonstrated in empirical applications on volatility and inflation forecasting. ",

author = "Jia Li and Zhipeng Liao and Rogier Quaedvlieg and Jia Li",

note = "Acknowledgments: We thank the Co-Editor (Francesca Molinari) and four anonymous referees for their comments and suggestions, which have greatly improved the paper. We also thank Raffaella Giacomini, Jinyong Hahn, Peter Reinhard Hansen (discussant), Hyungsik Roger Moon and conference and seminar participants at Aarhus, CREST, the 2018 Triangle Econometrics Conference, Southern California Winter Econometrics Day, the 2019 Toulouse Financial Econometrics Conference, and the 2021 SoFiE seminar for their comments. Liao{\textquoteright}s research was partially supported by National Science Foundation Grant SES-1628889. Quaedvlieg was financially supported by the Netherlands Organisation for Scientific Research (NWO) Grant 451-17-009. Publisher Copyright: {\textcopyright} 2021 The Author(s) 2021. Published by Oxford University Press on behalf of The Review of Economic Studies Limited. Funding Information: Permission by Surface Technologies Ltd. to publish the results of this work and the help of Mr. Y. Srur and Mr. A. Ronen in texturing the seal rings and running the experiments is gratefully acknowledged. The work was partially supported by the Israel Commerce and Industry Ministry and by the Fund for the Promotion of Research at the Technion. Publisher Copyright: {\textcopyright} 2021 The Author(s) 2021. Published by Oxford University Press on behalf of The Review of Economic Studies Limited.",

year = "2022",

month = mar,

day = "6",

doi = "10.1093/restud/rdab039",

language = "English",

volume = "89",

pages = "843--875",

journal = "Review of Economic Studies",

issn = "0034-6527",

publisher = "Oxford University Press",

number = "2",

}

TY - JOUR

T1 - Conditional Superior Predictive Ability

AU - Li, Jia

AU - Liao, Zhipeng

AU - Quaedvlieg, Rogier

AU - Li, Jia

N1 - Acknowledgments: We thank the Co-Editor (Francesca Molinari) and four anonymous referees for their comments and suggestions, which have greatly improved the paper. We also thank Raffaella Giacomini, Jinyong Hahn, Peter Reinhard Hansen (discussant), Hyungsik Roger Moon and conference and seminar participants at Aarhus, CREST, the 2018 Triangle Econometrics Conference, Southern California Winter Econometrics Day, the 2019 Toulouse Financial Econometrics Conference, and the 2021 SoFiE seminar for their comments. Liao’s research was partially supported by National Science Foundation Grant SES-1628889. Quaedvlieg was financially supported by the Netherlands Organisation for Scientific Research (NWO) Grant 451-17-009. Publisher Copyright: © 2021 The Author(s) 2021. Published by Oxford University Press on behalf of The Review of Economic Studies Limited. Funding Information: Permission by Surface Technologies Ltd. to publish the results of this work and the help of Mr. Y. Srur and Mr. A. Ronen in texturing the seal rings and running the experiments is gratefully acknowledged. The work was partially supported by the Israel Commerce and Industry Ministry and by the Fund for the Promotion of Research at the Technion. Publisher Copyright: © 2021 The Author(s) 2021. Published by Oxford University Press on behalf of The Review of Economic Studies Limited.

PY - 2022/3/6

Y1 - 2022/3/6

N2 - This article proposes a test for the conditional superior predictive ability (CSPA) of a family of forecasting methods with respect to a benchmark. The test is functional in nature: under the null hypothesis, the benchmark's conditional expected loss is no more than those of the competitors, uniformly across all conditioning states. By inverting the CSPA tests for a set of benchmarks, we obtain confidence sets for the uniformly most superior method. The econometric inference pertains to testing conditional moment inequalities for time series data with general serial dependence, and we justify its asymptotic validity using a uniform non-parametric inference method based on a new strong approximation theory for mixingales. The usefulness of the method is demonstrated in empirical applications on volatility and inflation forecasting.

AB - This article proposes a test for the conditional superior predictive ability (CSPA) of a family of forecasting methods with respect to a benchmark. The test is functional in nature: under the null hypothesis, the benchmark's conditional expected loss is no more than those of the competitors, uniformly across all conditioning states. By inverting the CSPA tests for a set of benchmarks, we obtain confidence sets for the uniformly most superior method. The econometric inference pertains to testing conditional moment inequalities for time series data with general serial dependence, and we justify its asymptotic validity using a uniform non-parametric inference method based on a new strong approximation theory for mixingales. The usefulness of the method is demonstrated in empirical applications on volatility and inflation forecasting.

UR - https://www.scopus.com/pages/publications/85126807916

U2 - 10.1093/restud/rdab039

DO - 10.1093/restud/rdab039

M3 - Article

AN - SCOPUS:85126807916

SN - 0034-6527

VL - 89

SP - 843

EP - 875

JO - Review of Economic Studies

JF - Review of Economic Studies

IS - 2

ER -

Conditional Superior Predictive Ability

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this