WP-ARAAR-2023 Percentile Weights Regression

Abstract

- In this brief note, we will review quantile and unconditional quantile models, and propose a new method of Quantile Regression (QR) called Percentile Weights Regression (PWR). It is important to recall that the main objective of these models is to estimate appropriate coefficients for a given percentile, based on the rank of the dependent variable or its predicted part. This is because the impact of a given explanatory variable on the outcome can be heterogeneous across different levels of the outcome. For example, the incidence of a social assistance program can vary depending on the level of wellbeing. QR models are useful in showing such heterogeneity.


Members

Project leader: Abdelkrim Araar

Project researchers:


Journal publication

No journal publications.


Working Papers

Title Modified Size Comments Recommendations
Exploring Heterogeneous Effects: Quantile Models and Percentile Weights Regression 2023-10-30 1.43MB 0 0

Policy Briefs

No policy briefs.


Final report

No final reports.


Proposal

No proposals.


Copyright © 2008-2026 PEP. All rights reserved.
If you have any question or if you need assistance, please contact: info@pep-net.org.