BITSS / AES-Roy-Majors

Replication files for "Selective Migration, Occupational Choice, and the Wage Returns to College Majors" (Annals of Economics and Statistics, 2021)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

AES-Roy-Majors

This repository provides replication files for the paper "Selective Migration, Occupational Choice, and the Wage Returns to College Majors" written by Tyler Ransom and published in Annals of Economics and Statistics in 2021. The paper is available here.

System requirements

  • R version 4.0.x
  • packages called at the top of Data/import_and_clean19.R and Analysis/analysis_all.R
  • A machine with at least 32GB of RAM, four processors and at least 50GB of hard drive space

Data

  • Data comes from IPUMS USA
  • In accordance to IPUMS licensing requirements, I do not post the raw data here
  • For a list of variables included in the raw data extract, see Data/usa_00045.xml
  • To obtain the raw data, you will need to create an account at IPUMS USA and then create an extract using the variables listed in Data/usa_00045.xml
  • Save this raw data file in the Data folder

Steps to reproduce the results

  1. Obtain data as described directly above
  2. Run Data/import_and_clean19.R
  3. Run Analysis/analysis_all.R
  4. Run the following files in Simulations/:
    • simulation.R
    • simulationCorrEps.R
    • simulationCorrEpsCorrEta.R
    • simulation1e3.R
    • simulation1e3CorrEps.R
    • simulation1e3CorrEpsCorrEta.R
    • and then cometogether.R
    • (See Simulations/README.md for further details)

WARNING: it may take up to 3 days for the results to complete. This time may be reduced if you have more RAM and more processors. (The majority of computation time is spent on bootstrapping standard errors; this time can be reduced by increasing the number of workers on line 313 of Analysis/table3splitBoot.R.)

About

Replication files for "Selective Migration, Occupational Choice, and the Wage Returns to College Majors" (Annals of Economics and Statistics, 2021)

License:MIT License


Languages

Language:PostScript 93.0%Language:R 6.1%Language:TeX 0.9%