crfb-tob-impacts

Social Security taxation reform analysis with PolicyEngine.

Reform Modeling Stop Sign

Before any paid CRFB reform Modal launch, read REFORM_MODELING_BIBLE.md. It points to the controlling command-center plan and progress ledger. The required production artifact is a full reform output H5 in durable storage for each (year, reform) cell.

Quick Start

The quick-start commands below are legacy local-development guidance. They are not the CRFB reform-modeling relaunch path and must not be used to launch paid Modal reform work. For reform modeling, use REFORM_MODELING_BIBLE.md.

# Setup environment with uv and Python 3.13
uv venv --python 3.13
source .venv/bin/activate
uv pip install -e .

# Generate policy impact data (takes 15-30 minutes)
python scripts/generate_policy_impacts.py

# Run Next dashboard
cd dashboard && bun install && bun run dev

Project Structure

• src/ - Core Python modules (reforms, calculations)
• dashboard/ - Next.js current-results dashboard
• paper/ - Quarto manuscript for citation and formal review
• data/ - Generated CSV data files
• tests/ - Test suite

Deployment

The production site is deployed on Vercel as a combined static build:

• dashboard at /
• citable Quarto paper at /paper/

When NEXT_PUBLIC_BASE_PATH is set, both the dashboard and paper are emitted under that base path.

Local production-style build:

python3 -m pip install -e .
cd dashboard && bun install && cd ..
./scripts/build_vercel_site.sh

That writes the combined output to .vercel-site/.

Data Generation

The legacy local project uses PolicyEngine to simulate fiscal and household impacts:

• scripts/generate_policy_impacts.py - Data generation using PolicyEngine simulations
• Output saved to data/ and dashboard public/ directories
• Historical pinned examples may cite older PolicyEngine-US versions. Do not treat those examples as the active reform relaunch contract.