No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Quantile Regression with Panel Data

    Good morning to everyone, I'm working on my thesis using a panel data and Stata 14.1. My panel embodies information for all the 20 Italian regions from 2004 to 2018. My independent variables are employment rate, average age, population growth rate and Gini index while the dependent variable is family poverty rate (all the variables are iid). I decided use quantile regression to understand if a given policy, that is designed to reduce poverty through an increase in employment, is really effective even if the poverty is structural. I'm using the model for quantile regression described by José Machado and Joao Santos Silva in their paper "Quantiles via Moments". My problem is that after the usage of "xtqreg", on stata, i would like to make tests on the residual and test the paper's assumptions but the classical commands that i know cannot be used with xtqreg. I noted that the "xtreg, fe" results are the same as the 0.5 quantile regression, so i tested the results obtained with the fixed effect transformation and they are robust but i suppose that this tests are valid only for the 0.5 quantile. Anyone has suggestions?

    Thank you in advance for your time.

  • #2
    welcome to this forum.
    Please read the FAQ on how to post more effectively. Thanks.
    One of the most useful recommendations included in the FAQ is to share what you typed and what Stata gave you back via CODE delimiters. This approach worths more than million words aimed at describing what's the matter with your data/analyses.
    That said, let's hope that Joao Santos Silva will chime in (but, please, as per FAQ again, do not post him or other listers privately. Thanks).
    Kind regards,
    (Stata 16.0 SE)


    • #3
      Dear Matteo DC,

      I am not sure which assumptions you want to test because I am not aware of any tests valid in this context (that is work in progress). Anyway, from the output of the command can construct residuals.

      Best wishes,