Open Access
February 2020 Two-step semiparametric empirical likelihood inference
Francesco Bravo, Juan Carlos Escanciano, Ingrid Van Keilegom
Ann. Statist. 48(1): 1-26 (February 2020). DOI: 10.1214/18-AOS1788


In both parametric and certain nonparametric statistical models, the empirical likelihood ratio satisfies a nonparametric version of Wilks’ theorem. For many semiparametric models, however, the commonly used two-step (plug-in) empirical likelihood ratio is not asymptotically distribution-free, that is, its asymptotic distribution contains unknown quantities, and hence Wilks’ theorem breaks down. This article suggests a general approach to restore Wilks’ phenomenon in two-step semiparametric empirical likelihood inferences. The main insight consists in using as the moment function in the estimating equation the influence function of the plug-in sample moment. The proposed method is general; it leads to a chi-squared limiting distribution with known degrees of freedom; it is efficient; it does not require undersmoothing; and it is less sensitive to the first-step than alternative methods, which is particularly appealing for high-dimensional settings. Several examples and simulation studies illustrate the general applicability of the procedure and its excellent finite sample performance relative to competing methods.


Download Citation

Francesco Bravo. Juan Carlos Escanciano. Ingrid Van Keilegom. "Two-step semiparametric empirical likelihood inference." Ann. Statist. 48 (1) 1 - 26, February 2020.


Received: 1 January 2018; Revised: 1 November 2018; Published: February 2020
First available in Project Euclid: 17 February 2020

zbMATH: 07196527
MathSciNet: MR4065150
Digital Object Identifier: 10.1214/18-AOS1788

Primary: 62M10
Secondary: 62G10

Keywords: empirical likelihood , high-dimensional parameters , semiparametric inference , Wilks’ phenomenon

Rights: Copyright © 2020 Institute of Mathematical Statistics

Vol.48 • No. 1 • February 2020
Back to Top