Overview
This project provides 10-year labor supply and demand projections (2024-2034) for AI-relevant occupations across 50 major U.S. metropolitan areas. Unlike traditional "labor shortage" analyses, we frame the question economically:
How many additional workers are needed to meet projected demand while keeping real wages growing at the rate of GDP per capita (~1.5%/year)?
A positive "gap" indicates demand exceeds supply—without additional workers, wages would rise faster than productivity growth. A negative gap (surplus) suggests potential downward wage pressure.
Supply Projections
Data Source
American Community Survey (ACS) microdata from 2013-2023, accessed via IPUMS USA.
Methodology
We use a cohort-based projection model that tracks workers by:
- Age cohort (5-year bands)
- Occupation (6-digit SOC codes)
- Metropolitan area (met2013 CBSA codes)
Entry and exit rates are estimated from observed flows in the 2013-2023 period. Supply projections assume these rates continue through 2034.
Demand Projections
Data Source
Bureau of Labor Statistics (BLS) Employment Projections Program, scaled to metropolitan area level using current employment shares.
AI Exposure Index
Felten, Raj, and Seamans (2021) occupational AI exposure scores, which measure the degree to which AI advances affect task performance in each occupation.
Automation Risk
Frey and Osborne (2017) computerization probability estimates, indicating the likelihood that an occupation's tasks could be automated.
Gap Calculations
The supply-demand gap is calculated as:
Gap = Projected Demand (2034) - Projected Supply (2034)
Where:
- Positive gap: Additional workers needed to prevent above-trend wage growth
- Negative gap: Potential worker surplus, possible wage pressure
Gap percentages are relative to 2024 baseline employment.
Scenario Modeling
We provide six scenarios with different demand multipliers by occupation category:
| Scenario |
Description |
| Baseline |
BLS projections without adjustment |
| Rapid AI Adoption |
High growth in AI/software roles; reduced demand for automatable tasks |
| Moderate AI |
Balanced AI integration across occupations |
| Slow AI Adoption |
Conservative technology adoption; traditional employment patterns |
| Green Economy |
Strong growth in clean energy and infrastructure occupations |
| Manufacturing Reshoring |
Increased domestic manufacturing and semiconductor production |
Limitations
- Assumes 2013-2023 entry/exit patterns continue through 2034
- Does not model inter-occupation mobility or geographic migration
- AI impact scenarios are illustrative, not predictive
- Metro definitions use 2013 OMB CBSA codes (met2013)
- Does not account for potential policy changes (immigration, education, etc.)
- Current data is MOCK for visualization development
Data Downloads
Download the underlying datasets used in this visualization:
Full Projections Dataset
Complete MSA × Occupation projections with supply, demand, and gap data
JSON • ~500KB
Download
Occupations Reference
AI-relevant occupations with SOC codes, categories, and exposure indices
JSON • ~25KB
Download
Metropolitan Areas
50 MSAs with CBSA codes, population, and regional classification
JSON • ~10KB
Download
Scenario Definitions
Demand multipliers for each scenario by occupation category
JSON • ~5KB
Download
Citation
Bahar, D. (2026). Labor Supply and Demand Projections by Metro and Occupation: A Cohort-Based Approach. Brown University Working Paper.
For questions or collaboration inquiries, contact: dany_bahar@brown.edu