ECON 499: Labor Unions and Racial Wage Inequality

PDF: https://aadsouza.github.io/econ499/Labor_Unions_and_Racial_Wage_Inequality.pdf

Contents

• /code
  • cleaning:
    • clean_nber_morg.do
      • cleans nber extracts of cps morg.
    • do_clean_nber_morg.do (in: morgXX.dta ; out: cleaned_nber_morg.dta)
      • do clean_nber_morg.do
    • pareto_topcoding.do (in: cleaned_nber_morg.dta)
      • gen top-code adjusted wage distribution using pareto parameters following fll2021.
  • pre-cleaning:
    • gen_cpi.ipynb
      • generates cpi using cpiaucsl from fred + other stuff from fred.
    • dind_nind_crosswalk.xlsx
      • crosswalk between cps industry and fll2021 industry categories.
    • state_cps_fips_abb_name.csv
      • adapt crosswalk from dm2020 for state, statefips, stateabb, state.
  • analysis (either in paper/appendix or referenced existence in replication package):
    • dfl1.do
      • heywood parent extended dfl decomposition.
    • linreg.do
      • state-ind unionization rate ols regressions.
    • linreg2.do
      • naive ols wage on union coverage.
    • med_wage_time.do
      • gen median lwage3 time trends by race and sex.
    • nindnocctab.do
      • percentage and union coverage in each ind table.
    • rifdid-plots.ipynb
      • rifdid plots w/o sltt (state linear time trends).
    • rifdid-sltt-bw-plots.ipynb
      • rifdid plots w/ sltt separately by race and sex.
    • rifdid-sltt-bw.do
      • run rifdid separately for B/w w/ state linear time trends
    • rifdid-sltt-plots-wo-ar2w.ipynb
      • rifdid plots w/ sltt w/o always rtw states.
    • rifdid-sltt-plots.ipynb
      • rifdid plots w/ sltt.
    • rifdid-sltt-wo-ar2w.do
      • run rifdid w/ sltt w/o always rtw states.
    • rifdid-sltt.do
      • run rifdid w/ sltt.
    • rifdid.do
      • run rifdid w/o sltt.
    • rifdiddiagnostics.do
      • run rifdid separately for B/w w/o sltt.
    • rifreg2.do
      • run naive rifols wage on union coverage.
    • rifreg2bw.do
      • run naive rifols wage on union coverage separately by race and sex.
    • rifreg2plots-bw.ipynb
      • rifols wage on union coverage plots separately by race and sex.
    • rifreg2plots.ipynb
      • rifols wage on union coverage plots.
    • stag_event.do
      • run event study that doesnt work.
    • stagdid-ptaplot.do
      • gen plots for predicted real log wage trends.
    • stagdid-wo-ar2w.do
      • run did - effect of rtw on B/w wage inequality - w/o always rtw states.
    • stagdid.do
      • run did - effect of rtw on B/w wage inequality.
    • stagsynthblack.do
      • synth - effect of rtw on wages for Black people.
    • stagsynthwhite.do
      • synth - effect of rtw on wages for white people.
    • stagsynthwhitewone.do
      • synth - effect of rtw on wages for white people. - dropped states for lack of common support.
    • sumstats.do
      • gen sumstats table.
    • ucov_race_sex_time.do
      • gen coverage rate time trends by race and sex >= 1983.
    • stagdid-fe-sltt.do
      • run did - effect of rtw on B/w wage inequality w/ year fe and state linear time trends.
    • stagdid-fe-sltt-wo-ar2w.do
      • run did - effect of rtw on B/w wage inequality w/ year fe and state linear time trends - w/o always rtw states.
    • med_centile_time.do
      • gen plot showing median person's position on white wage distribution.
  • miscellaneous stuff + graveyard (attempts and analyses that may not have made it into the paper)
    • restructure_mwage.do (in: mw_state_quarterly.dta ; out: qmwage7919.dta)
      • restructure Zipperer's quarterly state minimum wages.
    • dist_dataprep.do (in: cleaned_nber_morg.dta)
      • prepare data for distribution regressions following fll2021.
    • distreg1.do
      • first (only?) attempt at distribution regression following fll2021.
    • marg_crate.do
      • marginal effects of unionization rates consistent with fll2021.
    • bacondecomp.do
      • intended to run bacondecomp i.e. Goodman-Bacon decomp of stagdid (need to manipulate goodman-bacon ado to accomodate interaction).
    • cit_hisp_linreg.do
      • state-ind unionization rate ols for Hispanic people with citizen variable.
    • kde1.do
      • yearly kernel density estimates for white people, Black people, and Hispanic people.
    • morediagnostics.do
      • "more diagnostics"? for the RIF-DiD analysis of RTW laws (not sure what this ended up being).
    • rifdistributionplots-revised.ipynb
      • draft?
    • rifdistributionplots8919-bootstrap-w-errors.ipynb
      • i guess there are errors here?
    • riffig7.ipynb
      • more rifs.
    • rifreg1.do
      • rif on dfl attempt.
    • rifreg3.do
      • rifols state-ind unionization rate regressions.
    • stagarnr
      • compare alwaysr2w to neverr2w - so treat_st = treat_s (not DiD, post treat and post cont comparison)?
    • staghisp.do
      • run did - effect of rtw on Hispanic-white inequality.
    • dfl-3per.do
      • heywood parent extended dfl decomposition, only 8388 period.
• /dfl1
  • dfl1.log
• /distreg
  • note: distribution regression log files.
• /figures
  • note: figures and plots for analysis and graveyard.
• /linreg
  • note: log files for most analysis and graveyard.
• /tabs
  • note: tables for analysis and graveyard.
• lab_var_wage.txt
  • wage and related variable labels outline
• 495preliminaryprospectus.bib
  • bibtex file for most/all literature cited in proposal.
• 499data.bib
  • bibtex file for literature cited in thesis.
• thesis-defense-slides.pdf
  • beamer slides used for honors thesis defense.
• ECON_499_003__Honours_Thesis.pdf
  • final draft of thesis submitted to vse.
• Labor_Unions_and_Racial_Inequality.pdf
  • final pdf.

NBER extracts of CPS MORG

• econ499data directory with cleaned data etc. - used in .do files.
• NBER extracts of CPS MORG found here.
• NBER CPS MORG Documentation found here.

Acknowledgements

I am extremely grateful to Professor Nicole Fortin and Professor Marit Rehavi for their continued guidance and support. I thank Professor Thomas Lemieux for extensive feedback and insights. I am indebted to Professor Nicole Fortin and Neil Lloyd for generously providing Stata codes that expedited the data cleaning process. I also thank Sheldon Birkett, Felipe Grosso, Elisabeth Hatting, Wenxin Ma, Evan Mauro, Javier Cortés Orihuela, and Sarah Kirker Wappel for helpful discussions and comments. All errors are my own.