Comments on: A review on empirical likelihood methods for regression

Abstract

This paper provides an excellent overview and appraisal of empirical likelihood methods for regression. Different to earlier reviews on the empirical likelihood method, this overview covers quite comprehensively the more recent advances in semi- and nonparametric estimation and testing using this approach. Moreover, the authors consider the problem of inference with missing observations, censured and truncated data. All these topics have received considerable attention quite recently, and professor Song Xi Chen and Ingrid Van Keilegom are probably the leading researchers in this field. With this review they found a way to describe the—or better said, their—new advances in a quite friendly manner, without much theoretical and technical details. Although the title focuses on regression, the authors discuss further inference and testing as well. The only critical comment that occurs to me is that, while they highlighted very well advantages and benefits, they have been quite silent about drawbacks and obstacles.