Add SSTB QBI split inputs to us-data

[PavelMakarchuk]

Summary

Add the SSTB split inputs needed by the parallel policyengine-us QBID changes.

This PR exposes:

• sstb_self_employment_income
• sstb_w2_wages_from_qualified_business
• sstb_unadjusted_basis_qualified_property

in the PUF/calibration pipeline using the existing business_is_sstb flag.

Implementation

• Split self_employment_income into non-SSTB and SSTB pieces using the current all-or-nothing business_is_sstb indicator.
• Expose SSTB allocable W-2 wages and UBIA fields alongside the existing aggregate QBI guard-rail variables.
• Add the new fields to IMPUTED_VARIABLES so they flow through PUF-based calibration.
• Add targeted calibration-list tests and document the new variables in the appendix.

Important limitation

This does not infer mixed SSTB and non-SSTB allocations within the same record. The current data pipeline only has an all-or-nothing SSTB flag, so mixed-category wage/UBIA allocation remains approximate until more granular source data or imputation is added.

Verification

• PYTHONDONTWRITEBYTECODE=1 python -m py_compile policyengine_us_data/datasets/puf/puf.py policyengine_us_data/calibration/puf_impute.py policyengine_us_data/tests/test_calibration/test_puf_impute.py
• pytest -q policyengine_us_data/tests/test_calibration/test_puf_impute.py (blocked here because this environment is missing torch, imported by the repo test setup)

[MaxGhenis]

Requesting changes.

• GitHub reports mergeable=CONFLICTING and mergeStateStatus=DIRTY; a local git merge-tree --write-tree --name-only HEAD FETCH_HEAD confirms a content conflict in policyengine_us_data/datasets/puf/puf.py. That explains the missing CI on this PR.
• The PUF split moves SSTB Schedule C income out of self_employment_income and into sstb_self_employment_income, but SOI replication still builds business_net_profits and business_net_losses from only pe("self_employment_income") in policyengine_us_data/utils/soi.py. After this change, SSTB Schedule C profits/losses disappear from those total business-income comparisons. Please aggregate self_employment_income + sstb_self_employment_income anywhere the statistic is total Schedule C/self-employment income, or keep a separate total variable for validation.

Verified locally: uv run pytest policyengine_us_data/tests/test_calibration/test_puf_impute.py -q passed: 14 tests.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
pipeline-diagrams	Error		Apr 10, 2026 8:54pm

[MaxGhenis]

Approved after pushing the main merge/conflict resolution and SOI total Schedule C fix. Local verification: tests/unit/calibration/test_calibration_puf_impute.py and tests/unit/test_soi_utils.py pass.