We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: Inference in partially identified models with many moment inequalities using Lasso

Abstract: This paper considers inference in a partially identified moment (in)equality model with many moment inequalities. We propose a novel two-step inference procedure that combines the methods proposed by Chernozhukov, Chetverikov and Kato (2018a) (CCK18, hereafter) with a first step moment inequality selection based on the Lasso. Our method controls asymptotic size uniformly, both in underlying parameter and data distribution. Also, the power of our method compares favorably with that of the corresponding two-step method in CCK18 for large parts of the parameter space, both in theory and in simulations. Finally, we show that our Lasso-based first step can be implemented by thresholding standardized sample averages, and so it is straightforward to implement.
Comments: 1 figure
Subjects: Statistics Theory (math.ST); Methodology (stat.ME)
Cite as: arXiv:1604.02309 [math.ST]
  (or arXiv:1604.02309v4 [math.ST] for this version)

Submission history

From: Anders Bredahl Kock [view email]
[v1] Fri, 8 Apr 2016 11:19:03 GMT (155kb,D)
[v2] Mon, 25 Apr 2016 15:36:34 GMT (162kb,D)
[v3] Mon, 23 May 2016 17:44:59 GMT (155kb,D)
[v4] Sat, 29 Jun 2019 10:18:30 GMT (62kb,D)

Link back to: arXiv, form interface, contact.