AI-Era Skill Wage Premia: Hedonic Wage Regression Across Three Labor Markets

Hedonic wage regression on ~8,000 job postings across the UK, Canada, and Australia estimates wage premia associated with AI-adjacent skills in the post-GPT labor market.

IN PROGRESS LABOR ECONOMICSMLWAGE REGRESSION

About This Project

This is collaborative research with Dr. Anindya Sen at the University of Waterloo. I am the research assistant; Dr. Sen is the lead. Findings and methodology belong to the project and will be shared when the work is published and attributed.

My Contribution

  • Constructed and cleaned the job postings dataset (~8,000 postings across the UK, Canada, and Australia)
  • Designed data validation that caught parsing errors spanning five orders of magnitude — errors that, once corrected, materially changed point estimates
  • Built ETL workflows in Python (pandas, regex-based salary parsing) for reproducible cross-jurisdiction processing
  • Ran initial hedonic wage regressions to estimate skill-specific wage premia

The validation work was the substantive contribution: the raw data had systematic errors that would have produced misleading estimates. This is the kind of problem that doesn’t appear in a methodology section but determines whether the results mean anything.

Status

Research in progress. No findings shared before peer review and PI sign-off.