Leave-two-out cross fitting bias corrected test for cross-section dependence
Title: Leave-two-out cross fitting bias corrected test for cross-section dependence
DNr: SNIC 2022/22-1101
Project Type: SNIC Small Compute
Principal Investigator: Simon Reese <simon.reese@nek.lu.se>
Affiliation: Lunds universitet
Duration: 2022-11-11 – 2023-12-01
Classification: 50201
Keywords:

Abstract

This project aim at using a leave-two-out approach to bias-correct tests for cross-section correlation in residuals from linear panel data regression models. The leave-two-out approach addresses the problem of correlation induced by demeaning or defacoring data across cross-sections in a longitudinal/panel dataset. The computational resources will be used for monte carlo simulations that are too computationally intense for my desktop PC at work.