Inference based on Kotlarski's Identity
Speaker(s)Yuya Sasaki (Vanderbilt University, United States)
LocationUvA - E-building, Roetersstraat 11, Room: E5.22
Date and time
May 10, 2019
16:00 - 17:15
Kotlarski’s identity has been widely used in applied economic research. However, how to conduct inference based on this popular identiﬁcation approach has been an open question for two decades. This paper addresses this open problem by constructing a novel conﬁdence band for the density function of a latent variable in repeated measurement error model. The conﬁdence band builds on our ﬁnding that we can rewrite Kotlarski’s identity as a system of linear moment restrictions. The conﬁdence band controls the asymptotic size uniformly over a class of data generating processes, and it is consistent against all ﬁxed alternatives. Simulation studies support our theoretical results.
This is joint work with: Kengo Kato (Cornell) and
Takuya Ura (UCDavis).